ALUMNI NEWSLETTER
Spring / Summer 2016
Spring 2016
Chair’s Announcement
Dear alumni, students, colleagues, and friends: The excitement continues to grow around Berkeley Industrial Engineering and Operations Research. This year, we saw our ranking improve and we are now tied as the #2 ranked program in the US. Our students continue to be in high demand as more employers require highly skilled workers in order to compete in today’s marketplace. Dealing with data, optimizing processes, and leveraging computational tools are critical components of success for startups and large companies alike — and these are areas where our students excel. Further, as Silicon Valley and the Bay Area grow in global influence, we are teaching students entrepreneurial skills and working to improve the startup ecosystem on campus to create an environment where students can use their innovation skills to solve big problems. This spring, I read the names of 122 new graduates including nine ORMS students, thirty-two IEOR students, fifty-nine Master of Engineering students, fifteen Master of Science students, and seven PhDs. We are extremely proud of all of our graduates, and we can’t wait to see what they do next! At our graduation party/awards ceremony, we also recognized the achievement of students who went above and beyond with outstanding achievements in leadership and academics this year. (You can find a full list of awardees on the last page of this newsletter.) Thanks to the generous support of alumni, we also finished the first phase of our Etcheverry renovation this spring. Etcheverry’s 4th floor, the home of IEOR, has been truly transformed from aging to state-of-the-art. This has made a huge impact for students, faculty, and staff, and there is already enhanced collaboration and increased energy in the department. Inside the newsletter, you’ll find photos and a story about how the renovations have improved our environment. While this phase is complete, we hope to continue to work to update all of Etcheverry, including our departmental classroom on the first floor, so that the next generation of students can have the best opportunity to learn here. Alumni can help make this a reality by giving directly to the renovations campaign on our giving site here: https://give.berkeley.edu/browse/?u=325 We are also happy to announce that Paul Grigas will be joining us this fall as a new assistant professor. Paul recently received his PhD at MIT and brings skills and interests in convex optimization, machine learning, and data-driven decision making that will continue to grow the department’s ability to explore cutting-edge research issues, and help us educate our students on the latest theory and tools. This spring, we held an alumni mixer at ThirstyBear Brewing Company in San Francisco. Around seventy-five people joined us for the event, and we received lots of feedback that the attendees were excited to have the opportunity connect with fellow alums and faculty. We continue to look for ways to help alumni connect in the Bay Area and elsewhere. Please contact me or Keith McAleer (kmcaleer@berkeley.edu) if you have ideas or would like to host a future alumni social. Finally, it was a great honor for the department this year to learn that one of our faculty, Shmuel Oren, was elected to the National Academy of Engineering. Shmuel’s work has had a profound impact on the way that energy markets work in in the US and around the world, and it is great to see his work being acknowledged in this way. Inside this newsletter, you will find an exclusive interview with Shmuel. As always, please feel free to contact me at kaminsky@berkeley.edu with comments, questions, industrial project opportunities, or suggestions (and as always, donations are appreciated and can be made to a campaign of your choice by visting our online giving portal). Go Bears! Phil Kaminsky Professor & Chair, IEOR
IEOR Alumni Newsletter | FALL 15
Inside FEATURE • Shmuel Oren Elected to the National Academy of Engineering ...................................................................................................................................................................................... 4 ALUMNI • IEOR Alumni Profiles ................................................................................................................................................................................................................................................................. 6 • The Other Side of the Outlet ................................................................................................................................................................................................................................................... 8 STUDENTS • Alpha Pi Mu ................................................................................................................................................................................................................................................................................. 10 • Institute of Industrial Engineers ............................................................................................................................................................................................................................................ 12 • Students Connect Through ‘Lunch And Learn’ ................................................................................................................................................................................................................... 14 • Students Organize First Annual IEOR Graduate Student Symposium ............................................................................................................................................................................ 16 • Master of Engineering project profiles .................................................................................................................................................................................................................................. 18 SUTARDJA CENTER • Introducing the 2016 #WhatsNext Watchlist ...................................................................................................................................................................................................................... 20 • Berkeley Team Wins Geisinger Health Collider.................................................................................................................................................................................................................... 20 • Ben Horowitz Visits Newton Lecture Series ....................................................................................................................................................................................................................... 22 • Engineering Leadership Alumni Gather at Samsung’s New HQ for Mixer ..................................................................................................................................................................... 24 • Building a New Bridge to Silicon Valley Culture ................................................................................................................................................................................................................... 25
Shmuel oren elected to the national academy of engineering
4 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
This year Shmuel Oren, the Earl J. Isaac Professor in the Science of Analysis and Decision Making, was elected to the National Academy of Engineering (NAE) for his contributions ‘to the integration of decisions and cooperative market mechanisms for adaptive multisource electrical power systems.’ Shmuel was one of three Berkeley engineers elected to the academy this year and one of eighty worldwide. Below is a transcription of an interview with Shmuel which has been lightly edited for print. IEOR: How have energy systems and markets changed since you have come to Berkeley? Shmuel Oren: I spent the first ten years of my career working on theoretical optimization, but when I started working at Xerox Palo Alto Research Center I became interested in applying optimization to marketing, pricing, and product line design. My interest focused on the use of markets and economic principles from an engineering perspective — I actually got my degree in Engineering-Economic Systems. The basis of my work has been to think about economics the way most engineers think about physics as in understanding the relationships between entities, and then taking the next step and asking: how can we design mechanisms that can harness the forces of greed, profit-seeking, selfish-preservation or whatever you want to call it to guide the system if we had perfect information and were all powerful? The whole idea of market design and market engineering — which I’ve been trying to promote — is to design a set of rules that will provide incentives to the participants to take actions out of selfish motives that would result in the solution you would want to prescribe. When I was at Xerox I was focused on the pricing of telecommunications networks and on how to create product lines and product differentiation. Very soon after that in the ‘80s — especially during the first energy crisis — I got interested in how to apply these ideas in the context of energy. I was working with EPRI, the Electric Power Research Institute, and together with Professor Wilson from Stanford we started a project focusing on the idea of priority service and efficient rationing — how to, for example, design a product line for energy service that will give people the choice to pay more and get higher reliability or higher quality of service or pay less to save money — which is a common idea in many other areas. When you buy a car or anything else there are choices. You buy a first-class seat in a plane and you pay more or you buy a coach seat and you pay less. Our idea was to introduce these concepts into the energy market. Now it is being called demand response. Currently, we have smart meters, time-of-use rates in our homes, and even some real-time pricing which for example might show a customer how much it costs to charge their car at night versus during the day. This induces the customer to opt for the lower price unless there is some urgency. The general idea is to give people the correct price signal of how much it really costs to serve them, so they have the same opportunity as they would in a supermarket where they have choices and decide how much they want to spend. In energy, this was never the case — your meter was read at the end of the month, and you were billed accordingly. So, this idea of introducing more product differentiation and market segmentation in energy launched my energy career. Then came the re-structuring of the energy industry where we moved away from vertically integrated utilities like PG&E that owned and controlled all the generation, transmission, and distribution of energy. We moved to a more competitive market where now PG&E acts more like a buyer on behalf of consumers, and there are regulated rates determining how much they can charge you. But they can still offer deals to some customers. For example, if you are willing to be interrupted a certain number of times during the week or during the month when there is a hot day, you can get a discount on your bill. Those were the kind of ideas that I had been exploring in the ‘80s, but they were never implemented on a large scale. PG&E now doesn’t own the generation anymore. They either buy the energy through long-term contracts or they buy it through an exchange, like a commodity exchange, where the generators offer quantities and prices based on fuel price and availability. Then the system operator is running an auction in the day ahead for the next 24 hours and the price fluctuates depending on supply and demand every 5 minutes. These are the prices PG&E sees but they buy most of the energy in advance in a forward market, so they are protected against high volatility. If you are a big customer, you can buy energy directly from the system operator at wholesale prices. IEOR: What are the hottest topics in energy systems today? SO: The hottest topics these days have to do with renewables integration. Renewables bring a whole new set of elements, because when you start to have a significant portion of the electricity supply coming from wind and solar, it is no longer controllable. Sometimes, the wind blows and sometimes it doesn’t blow. So, the challenge is: how can we operate a reliable electric power system where there are several generators that each take awhile to come online and at the same time account for all of the uncertainty and volatility of those renewable resources. We have to keep enough energy resources available to meet demand when the wind stops blowing or a cloud blocks the sun, but if we are always keeping spinning reserves online to be ready when there is no sun or wind then that is going to undermine a lot of the economic and environmental benefits that renewables provide. For one of the projects that we are involved in, we are working to arrange contracts for customers who are willing to occassionally reduce their load and allow the utility or an aggregator to cap how much electricity they can draw. With wind for example, you will get cheaper electricity, if you are willing to shift some of your load by doing laundry or charging your electric car at night. So, in exchange for lower rates, the customer gives the utility permission to limit their usage (or face a penalty). This empowers the customer to make those choices and at the same time the customer acts as a backup for the wind uncertainty. This is a new service paradigm. In a traditional electricity system generators follow the load whenever the customer turns the switch on or off. Now customers can choose to follow the energy, whenever it is available (and at a reduced rate). There are other aspects to this new energy system transformation. The first challenge is designing the system from a technical perspective to make sure that energy supply matches demand. To make this work, you have to create economic incentives so that people will behave the way you actually want them to behave. Beyond designing incentives to guide behavior we also are using computer simulation to try to understand how the system is going to shake out when you are actually operating it. In our theoretical analysis, we view the market as a game where we are making the rules of the game and everyone acts accordingly. So, this is the new way of optimizing the system by creating a set of rules where everyone in a distributed fashion can manipulate whatever they can control, and we end up with the net result which would be equivalent to what you would have done if you had all the power in the world to tell people what to do and all the information you needed to control the system. In short we are trying to reconcile the availability of information, the availability of control, and the fact that everybody is a player. To do that, we are taking advantage of the fact that we have all these new computational resources. For example, I have a student that worked on algorithms to run the system where we not only control how much every generator is going to produce, but we can also switch lines in and out of the transmission networks. So, essentially we are controlling the topology and generation of the network simultaneously. This enables us to be like a traffic controller where we can re-route some of the electricity and use cheaper resources that we would otherwise not have access to. There are many issues at the interface between engineering and economics and because you also have to plan ahead, manage risk, and sign contracts in this type of environment, we are also employing risk management and financial engineering tools as well. (continued on page 26) (continued from page 5) For example, PG&E signs contracts, but those contracts have to be hedged against fluctuation in fuel supply, and the price of gas, so a lot of the financial engineering concepts are working their way into commodity hedging to help us understand and manage risk. The OR tools, optimization, economics and risk management all come together. That’s why my students end up taking 2-3 classes in the economics department, classes in risk management, electrical engineering, and operations research. I have several students that are working at PG&E. One of our graduates Todd Strauss, who is also on the IEOR advisory board, is now a Senior Executive Director at PG&E in charge of regulatory affairs. I have students that end up in electrical engineering departments, some that ended up in economics departments or operations research depending on their particular interest and orientation. Our energy program has cast a big shadow in this respect. IEOR: In the future will prices update in real-time for customers? SO: In the old days they talked about rate-payers. Basically, they would tell you how much you have to pay and you pay it. We no longer talk about rate-payers; we talk about consumers. In fact, people talk about prosumers, because every consumer is potentially a producer. If you have solar panels on your roof, you also produce electricity, so you are a prosumer. And if you have a storage device at home or an electric car, you are a prostumer — a producer, storage manager and consumer. So, people talk more and more about decentralizing the supply of electricity. For instance we can have microgrids — such as at a university — with only few points of contact with PG&E. Much of the load balancing is internal within the subsystem (e.g. within the campus), and this reduces reliance on the transmission grid as a whole. All these new concepts on how to re-organize and de-centralize some of the supply is what makes this whole area so fascinating for operations research and industrial engineering. Electricity is probably the most complicated supply chain in the world. If you think of it from a supply chain perspective, first of all you have this situation where supply and demand have to be balanced on the fraction of a second. You have very little storage — storage is very expensive — so you don’t have inventories that you can use to smooth out production. You have congestion in the system that create a lot of routing/transportation problems. You have a lot of uncertainty because the supply of renewables is highly variable. So, we face all these challenges that people struggle with in the supply chain context where the time scales are much larger, but here you have to do everything instantaneously. How much reserves to keep and how to balance the load become very challenging questions. Then you also have the economic aspects because this is an area where prices are actually reflecting cost. There is no other area where you have such instantaneous feedback to the consumer on how much it costs you to produce something. In electricity markets, like in California, the price of electricity changes not only every 5 minutes but also varies geographically among 3,000 locations. So the cost of consuming a KW in Fresno is very different than consuming it in San Francisco. Electricity markets represent the biggest application of optimization because the system is being optimized continuously. They have to employ a lot of OR people that are continuously improving and refining this optimization engine and the prices are coming directly out of the optimization process — telling us exactly how much it costs at every location. Then there is of course the whole area of gaming, because one of the problems is that the industry is so concentrated that a lot of the producers have market power, so they could withhold some capacity or pretend that a plant is down [to artificially inflate prices] — and that is what happened in 2000 during the energy crisis; we saw a lot of gaming — the market was not designed properly, so people were scheduling transmission much more than was actually needed and then they were charging the system operator to essentially withdraw some of the transactions. For instance a trader like Enron would schedule a transfer of 1000 MW from north to south, and the system operator would say, “well that is going to congest the line” because if the lines transfer too much power, they overheat, they start to sag, hit trees, create shorts, etc. So Enron would say, “okay, you have to compensate me, because if I cannot send 1000 MW from the state of Washington to Los Angeles it means that I have to buy the power in Los Angeles at a much higher price to serve my customers, and I have to withhold or sell back the power and put the water back in the reservoir.” This is a legitimate concern and is one reason why prices are so different from north to south. However, because of the way the market was designed, they were able to make such claims even when they didn’t have the transaction, because it didn’t cost them anything to over schedule in the day ahead. Then in real time they were collecting money to basically relieve congestion which shouldn’t have been there in the first place. So, that’s how Enron played the market. There are all kinds of games that can be played and you have to look at it from the point of view of strategic interaction and game theory and either do some simulation or mathematical analysis to understand the possible gaming strategies that the different players can have and how to design a market in a way that prevent the abuse of market power. We have a new thing on our plate with cap and trade and the whole issue of environmental regulation. We have limits on how much CO2 you can put in the air, on generator output and so on. For instance, we are dealing with something called shuffling. If we are going to put a limit on the CO2 plants in California, is that really going to make a difference or is it just going to be paper shuffling so that the generators in Nevada are still going to pollute, and we are going to pay more for electricity. If all we doing is shuffling contracts then we only impose extra costs on the people in California with no reduction of emissions. So, you have to analyze all these things to understand how imposing a carbon tax or cap and trade affects the equilibrium. Suddenly, you will find a plant that was originally considered cheap is now expensive because it has to pay for the carbon it produces, and another plant that was relatively expensive because it was clean, is now cheap, so that’s going to cause a whole new reconfiguration of the equilibrium and flow. So, we are working to understand what the interaction will be between the CO2 market, the renewables, the production of electricity and the prices in different places and how to design the market to make sure things work the way we want them to work. IEOR: Will cap and trade policies help reduce carbon emissions? SO: Ideally, the best solution would have been a carbon tax based on emissions. The tax should approximate the relative environmental damage that is caused with the carbon, and that would have regulated consumption and balanced things. The problem is that no one trusts that the government will use the money they collect from the tax for a good purpose, so then the question becomes very political. What cap and trade has tried to do is create a market solution. If we know how much carbon we want to allow, we can allocate permits for that amount and allow people to trade them amongst themselves. Some might find it’s cheaper for them to reduce their emissions by innovating their plant and pay for some of that cost by selling some of the permits they got to someone else that prefers to continue to burn dirty fuel. The idea is that the permits are going to go wherever it is most expensive to reduce the emissions and where it is relatively cheap to renovate the plant, people will find it more economical to sell their permits and reduce their emissions. This provides some flexibility, and the price of those permits will be determined based on the opportunity costs — what it really costs to eliminate or reduce the pollution. There are all kinds of things you can do: take clunkers off the market, pay people in permits, but eventually the idea is that you keep reducing the permits that are available and let the market establish who the next in line to make the investment is. This way the money stays in the pockets of the generators, and instead of having the government collect the tax and decide winners and losers by doling out money to improve the technology, you give out permits and say okay, this is the total pollution that we can anticipate and tolerate, here are the permits based on some historical usage and now you guys trade it among yourselves and figure out who is willing to pay the highest for the permits. The Cap and Trade idea would have also worked internationally. The problem with doing it internationally is the fear that the US is going to buy all the permits from China, and then the Chinese will have to reduce industrial development. We have already done a lot in terms of reducing our emissions, so we are at the steep part of the curve where it is very expensive to innovate. So, the low-hanging fruit is still in India and China. So, we can buy the permits from them and give them some cleaner technology and that will have a big impact on international trade. That’s why people aren’t jumping on the international Cap and Trade idea right way. But from an economic point of view that would have been the ideal solution. Within the states we have now a situation where California has the tightest controls on emissions and one of the issues that economists are concerned about is the so called ‘leakage’. Basically people in California used to buy power from the coal plants in Nevada or Arizona. Now we are imposing restriction on import from certain kinds of resources, so it forces the California consumers to buy power from gas or from resources that are more expensive. For instance instead of Californians buying cheap power from a power plant in Tijuana, they buy the power from plants in San Diego. However, the plant in Tijuana doesn’t necessarily produce less, they just sell the power to someone else. The net effect — this is where the leakage is — is that you don’t really change emissions; the impact is on the contracts and who gets the good deal and who gets the more expensive deal — and that’s the problem. So, that’s why we always have to monitor and analyze how the equilibrium is going to shake out when you impose these kinds of rules. It may be a good rule, but you have to take into consideration that you are dealing with clever players whose strategic responses may circumvent the rules. You cannot expect to run an economy or system — and that’s one of the principals of market engineering — trying to design a system based on altruistic behavior, which is like trying to design an airplane based on levitation. Greed and profit-seeking is like gravity — it’s there and cannot be assumed away. So, the objective of market engineering is to design systems that will harness the greed to achieve good outcomes, rather than assuming that people will do the right thing. There is a certain amount of good will, I have to say. People are buying electric cars and trying to pollute less in California just to be green and out of good behavior, but that only happens up to a point. Ideally, what you want is to have an economic systems that incentivize the behaviors that you expect, so people don’t have to feel singled out and end up spending money when their neighbor is free riding. IEOR: Can the grid be powered by 100% renewables? SO: In order to do that, you need a system where people are willing to put up with the fact that their lights are going to be out for a certain fraction of the day. There is a limit to how much renewables we can absorb, but we can try to do it in a way that will maximize their use. The game changer that will enable us to run the whole system on renewables is some incredible new technology for storage. Then we can store the energy during the plenty and use it at night or whenever we need it. So, storage can enable us to run a system just off renewables. In Denmark, for instance, they are targeting for 100% renewables because across the water in Norway they are powered by 97% hydro, so any time they have extra wind they pump water uphill, and every time they need energy they just use the hydro. So, you have a system that can store all the excess energy during the off peak hours and give it back to you when you need it. But when you have a system that relies on thermal resources where you do not have that much storage or that much hydro, then there is a limit of how much renewables you can integrate. But we can do it in a clever way. The place where our discipline comes in, is in the use of information and computation in order to better control the resources we have and optimize their use. IEOR: How much of the energy that is currently generated is wasted before reaching the consumer? SO: We don’t waste much [energy], but we waste money in the sense that if you wash your clothes when everybody is eating dinner and suddenly you have the peak, there is a lot of waste since you need to build more power plants in order to cover that peak. If you can help fill the gap, then you have more opportunity to flatten the load profile and reduce capacity needs. There is also a phenomena we call the “duck curve” which basically says that if you are looking at solar energy — and we are going to have more and more solar energy — then between sunrise and sunset a lot of our load will be to be served by solar energy, so the net system load shows a big dip, which is like the belly of the duck. Consequently at the beginning of the day and at the end of the day, we will see a sharp drop and sharp rise in net load which means that suddenly those conventional generators that are producing electricity have to ramp down quickly when the sun rises, and up to fill in when the sun sets. These ramps create tremendous challenges for traditional generators so more and more we need flexible generators that have the capability to up and down very fast. We are talking about 30% of the load suddenly you have to jump up and go down over a short period. That requires more expensive technology and exactly how to do that and manage this kind of technical constraint is a big challenge. My interest is always how to do it with market mechanisms, how to create incentives which will mobilize a lot of independent resources that will behave in certain ways in response to those price signals. The alternative is way too expensive. IEOR: Are there laws that make market based mechanisms difficult to create? SO: The laws are creating interesting outcomes. For example, just recently the California Public Utility Commission had to make a decision on whether to impose a connection charge, because the way it is set up now you just get charged for the energy you use. This is problematic; it worked in the old regime because people paying higher rates end up carrying the burden for all the infrastructure. What is happening now, is that all those people on the high tier rate are installing solar panels on their roof. There is something called net metering, where you not only get free electricity from the sun, but when you push back electricity into the grid, it turns the meter backwards and you end up getting paid for the excess energy that you are producing. So, this net metering has creating an environment where some people with big houses are not paying anything for electricity because they are on average zeroing out their electricity bill. But they are the ones we were counting on to cover the cost of the infrastructure. So, now the infrastructure will be paid for by the poor people who do not have solar panels. The CPUC was attempting to change the law to impose a connection charge so that no matter how much solar you have, you pay maybe a fixed connection charge such as $80 a month to be connected while the energy charge would be lowered. So, if you are saving on the energy, you don’t save as much as you did before. Because of politics, the CPUC voted 3-2 to continue net metering for another 5 years. Nevada went the other way. The biggest lobbying against discontinuing the net metering came from the solar industry. They argued that discontinuing net metering would put them out of business, and they won by one vote which makes purchasing solar right now very attractive. I immediately ordered solar panels. I was holding off because I was sure they were going to stop net metering and establish a connection charge. But now, I put in 6kw of solar on the roof that’s going to pay for most of my electricity and pay for the panels within the 7 years and have free electricity afterward. So, these are the kind of rules and regulations you have to understand because they create incentives and people behave differently depending on the incentive you create. In this case, it’s kind of trivial observation, but there are a lot of situations like this. IEOR: Can you tell us more about microgrids? SO: The idea is that you can have an industrial park or campus or similar complex where you work to balance the load internally as much as possible before you going outside of campus to draw from PG&E. There is a whole movement in that direction. New York now has a big initiative called REV (Reform Energy Vision) where are de-centralizing the grid and enabling some of the sub-systems to have their own balancing before tapping in to the main grid. Again it would be based on incentive, design, and it is how we envision the future of energy. It encourages renewables because there are incentives to balance the load locally. One of the big costs of energy, in addition to the consumption of energy, is the infrastructure, and the reserves which must be kept on standby to keep the lights on. More of this cost can be transferred to the micro-grid. If you are in a big high-rise building in New York, you have a controller that can manipulate the air-conditioners in the building and maybe put ice on the roof in the morning and recycle some of that hot air over ice and reduce consumption. So, there are a lot things that we can do locally to balance the load given the proper incentives as opposed to everybody turns the switch on and plugs in their electric car at the same time and the system operator has to keep enough resources to serve any foreseeable load which is very expensive. IEOR: Will something like the Tesla battery be a game-changer in terms of storage? SO: Tesla is promising to have cheap batteries. I don’t think they can achieve their goal in terms of cheapness. There are problems when you start to use these kind of batteries on a mass scale. You have to consider whether there is there enough lithium in the world to build those batteries and the environmental impact of all these batteries that they will have. There are always challenges. Those are the kinds of issues people have to think about. You can’t just solve a mathematical problem, you have to think of what the big picture is. We use technical tools to help us answer questions and design systems, but we have to develop domain expertise to understand what we are solving. Math is a good way to address these problems, but solving the right problem is also important. ■
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 5
This award was presented to Shmuel to acknowledge his work as part of the California Independent System Operator Market Surveillance Committee
Alumni
Irene Zarco Campero M.S. IEOR ‘13 Currently: Materials Program Manager at Amazon Lab126 What was your first job ever? Inventory Analyst at IBM (In Guadalajara, Mexico) What did you do after graduation? Worked as Supply Chain manager at Cisco for 8 months. Then I started working at Amazon Lab126 as Materials Program Manager. What are you currently working on? I manage all the materials aspects including schedule, cost, inventory, for new product introductions during the development phase until launch to market. I also work in developing the supply chain for new products, selecting suppliers and vendors and work with them during the development/design phase. What do you miss about being a student at Cal? Being around very intelligent people from very different backgrounds. Do you have any advice for current IEOR students? Read and get involved with different industries to understand how IEOR is applied differently depending on the industry.
6 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
Gemma Berenguer Ph.D. IEOR ‘12 Currently: Assistant Professor at Purdue University (Krannert School of Management) What was your first job ever? Picking up peaches at my grandpa’s orchard What did you do after graduation? I went straight to my current position as a tenure-track assistant professor in Operations Management. What are you currently working on? I am currently doing research on nonprofit and sustainable operations management and the design of global health supply chains. I also teach in the MBA program. Do you have any advice for current IEOR students? I miss the great courses that I could take and the overall environment on campus and its surrounding area. The good weather is also greatly missed :) What do you miss about being a student at Cal? Besides studying, make sure to communicate with your classmates and professors. They are the biggest asset of Cal!
IEOR Alumni Profiles Check out what alumni are currently up to
Bowen Nie M.Eng. ‘12 Currently: Worldwide Supply Demand Product Manager at Apple What was your first job ever? I worked as an Aggregate Planner in a Semiconductor Company called Xilinx in San Jose. What did you do after graduation? Took my parents on a wonderful road trip. Did a summer internship in China Development Bank in Beijing and then went back to the States to start my career. What are you currently working on? I’m currently working in Apple’s giant supply chain department. Managing end-to-end supply demand planning for one of the company’s hero products. That’s all I can talk about. What do you miss about being a student at Cal? RSF, Library. And food on Durant Ave. Do you have any advice for current IEOR students? Make friends and build up your network. Then try hard to expand that to include interesting people who are going to motivate you and to help you improve.
Farzin Shadpour B.S. IEOR ‘05 Currently: Leader, Supply Chain at Theranos Inc. What did you do after graduation? I joined the Boeing company. I got to work at Boeing’s main factory, which is the largest building in the world on the 747 program. Within 15 months I became lead-engineer, which is a record at Boeing, and three months after that moved on to the Supplier Management organization where I got the opportunity to work internationally for 4 years. What do you miss most about being a student at Cal? The opportunity to learn from amazing faculty and students, being in classrooms and labs, and participating in various student organizations and clubs. Do you have any advice for current IEOR students? Learn as much as you can. Do take your course projects seriously and treat them as a real-world problem within a corporation. Get internships. Reach out to Alumni and ask for mentorship and informational interviews.
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 7
Interested in sharing what you are up to with fellow alums? Take the 2016 alumni survey by visiting: svy.mk/1TMdgSi
Brian Zheng M.Eng. ‘14 Currently: Consultant at Deloitte What was your first job ever? HSBC banker What did you do after graduation? I did forensics analysis for two months before joining Deloitte consulting in SF. What are you currently working on? Government project. What do you miss about being a student at Cal? The intense learning and all the classmates. Do you have any advice for current IEOR students? Efficiently manage your time and choose wisely.
Lei Xu B.S. IEOR ‘11 Currently: Co-founder at Emote What was your first job ever? Associate, ZS Associates What did you do after graduation? Management consulting at ZS Associates. Then, strategy and operations at Google. Then, starting my own company Emote. What are you currently working on? I co-founded Emote in 2015 to help teachers better understand and support students. What do you miss about being a student at Cal? The diverse, intellectual discussions on a wide range of topics, as well as the ability to pursue divergent projects on a whim. Thinking back, having the freedom and bandwidth to think, challenge, and learn is truly amazing! Do you have any advice for current IEOR students? One mentor gave me this advice that I found apt in our 20’s: “take the time to figure out what you are meant to do in your 20’s, and you will have the next 30 years in your career to do it well.” I take that as optimizing for learning and growth whenever possible.
The Other Side of the Outlet Recent PhD graduate Paula Lipka joins PG&E’s short-term electric supply team
8 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
“We forget just how painfully dim the world was before electricity. A candle, a good candle, provides barely a hundredth of the illumination of a single 100 watt light bulb.” – Bill Bryson The electric world is an amazing one: illuminated screens bring us all the information that humans have compiled thus far on demand, large metal boxes hold our food and keep it cool so that it lasts longer, and outlets charge batteries for our mobile devices which we can then bring anywhere to communicate instantly with loved ones on the other side of the planet. While using power is a relatively painless act for the consumer, there is a frenetic pace of activity happening on the other side of the outlet. “The way you get your lights on here is through a market that figures out the schedule for all the generation, and then they manage the minute-to-minute fluctuations of load and production,” said Paula Lipka, a recent IEOR PhD graduate who has just begun working on the short-term electric supply team at PG&E. In order to keep prices low and meet ever-rising energy demands, providers such as PG&E must bid their energy resources (i.e. hydropower, gas, solar, etc.) into the competitive California Independent System Operator (CAISO) market. In her new position, Paula supports the day-ahead group as well as the real-time group which updates bids every 5 minutes (24 hours a day, 7 days a week). Because there are currently not many options for efficiently storing energy that has already been generated (though, Paula tells us PG&E is in the process of greatly increasing this and currently has a large pumped storage hydro-plant and couple of large batteries), in most cases any power generated must be used immediately by the consumer. And with renewables becoming a larger part of the picture, not only do energy providers have to competitively bid their resources into the market, but also account for more unreliability in the system, “You will have cases — especially in real-time — where a cloud [covers solar panels] and power goes out. There are also issues with over-generation, and there is a pretty huge ramp that occurs in the evening when solar goes away and everyone is coming home. There are also more price spikes and negative prices. Negative prices mean that there is too much energy in the market and instead of getting paid by the market, you have to pay to generate power into the market,” said Paula. So, for example, in the mid-afternoon when everyone is back from lunch and before the dinner-time rush of energy, there may be negative prices because more power was generated than needed (if the day was sunnier than expected, for example). So, one can see why when asked what the most important skill she learned while at Berkeley IEOR was, Paula did not hesitate before saying “optimization under uncertainty.” Paula is still learning about how things work at PG&E, but hopes that she can contribute to the process by updating the models to better account for uncertainty in its bids (i.e. how confident are we in our bid?), update the way data is visualized by the real-time team, and on a broader level help facilitate communications between the real-time and day-ahead teams. Of course, whatever models she comes up with, they will have to work in a matter of seconds to be relevant for the real-time team where strategies change every 5 minutes. Energy wasn’t always bought and sold in such a frenzied manner. Before the passing of the Federal Energy Policy Act which introduced competition into the wholesale energy market in 1992, the industry basically operated as a monopoly with one local utility typically in control of generation, transmission and distribution services for the area. Long-term deals were typically communicated by phone call and were not very market-based. This opened up the market for manipulation by industry. For example, in the past, providers could withhold power generation in order to artificially inflate prices. While it is still possible for providers to manipulate the market, it is much more difficult because the market is competitive (with prices changing every 5 minutes) and there are controls in place to help prevent market manipulation. “There are a lot of different strategies. They will do a number of different tests such as the pivotal supplier test which looks to see if anyone is dominating the market. If you are the person setting the price, and they see that your price is for some reason way higher than your default bid, then your bid will be mitigated to the default price,” Paula said explaining one way that the market works to prevent cheating. Clearly, the energy markets are one place optimization techniques are very valuable. With IEOR students being free to apply engineering tools in many different disciplines, Paula said she chose energy systems because she knew she wanted to work on practical problems and liked energy because there was so much real data available. She also knew the problems were very relevant and saw the potential to make a big impact as there is still a lot of work to do to sort through the various policies and approaches to making the market fair and optimal. Since picking a topic to focus on in IEOR can be challenging because there are so many possible fields where IEOR techniques are valuable, Paula recommended that students start by working to make connections in the areas they are interested in, “That can help a lot if you are a PhD student trying to come up with an interesting research topic. It’s often a lot easier and more fun if you are more aware of what is going on in the field you are interested in.” As far as how to get started making connections — which is not always easy — Paula said she was fortunate that her advisor Professor Oren helped her get an internship at the Federal Energy Regulatory Commission (FERC). She believes that internships are a great way to connect with industry because it forces you to connect with people, and because you are working on real projects together, it makes it easier to make those connections last. Though, Paula is excited to apply her knowledge and skills she knows she will definitely miss Berkeley IEOR. “There was a good group of people there and that was nice to have. Also, I have quite a bit of flexibility at my job, but not as much as a being a grad student. Having that super-flexibility to look at anything you want is something really useful — but it can also be challenging. So, I will miss that. Also, I will miss having such a wealth of knowledge available. Being able to learn about the best approaches that are going on right now is very useful.” So, the next time you plug in your iPad, laptop, or any other of the myriad devices that use power, remember Paula and the invisible market that works 24 hours a day to deliver reliable power to our outlets. ■
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 9
Alumni Resources CAL ALUMNI NETWORK - Create a profile and connect with other alumni. CAL ALUMNI ASSOCIATION - Join or create an alumni chapter, join events + volunteer and travel opportunities. ALUMNI CAREER SERVICES - Alumni career resources + opportunities to help students. CAL REUNIONS - UC Berkeley automatically organizes reunions every 5 years. ENGINEERING ALUMNI NETWORK - Update your information here and receive invites to special events, the Berkeley Engineer magazine, connect with other alums and more. BLUE AND GOLD YEARBOOK - Re-order most Cal yearbooks since 1997
“Having that super-flexibility to look at anything you want is something really useful — but it can also be challenging. So, I will miss that. Also, I will miss having such a wealth of knowledge available. Being able to learn about the best approaches that are going on right now is very useful.”
Mehran Navabi Treasurer My past three years at UC Berkeley have allowed me to explore my interests and passions while surrounding myself with some of the most interesting and motivated people I have ever met. I have also gained experience in a variety of engineering fields, including embedded systems programming and data analytics. APM has provided me with outstanding professional guidance and resources, as I continue to search for new opportunities to share and apply the skills and knowledge I have gained at Cal with the rest of the world.
To see the latest news from Alpha Pi Mu at UC Berkeley, visit: apm.ieor.berkeley.edu
10 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
Current Students
Bowen Ni Vice President I plan to change to world to a more efficient place. In term of future career, I am interested in data analysis, supply chain, product management and operations. One of my favorite thing I learned is that machine learning is really cool. My favorite thing about being an APM member is that I have a great group of members that I can help me to explore professional career channel.
Alpha Pi Mu Alpha Pi Mu (APM) is the honor society for individuals who have shown exceptional academic interest and abilities with industrial engineering and operations research. Below are current officers at Cal.
Aaron Prohofsky President Everyone I meet asks me what industrial engineering is. To me, industrial engineering takes the technical skills of an engineering field and applies them to make optimal business decisions. Data analytics, supply chain management, and consulting are all great fields for an IEOR major. I am still undecided in terms of my career path, but I do know that an IEOR degree has taught me many tangible skills, along with the ability to problem solve and think critically. I can’t wait to apply them in the professional world. APM is great because it allows students with common academic and career interests to form a social bond!
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 11
Alpha Pi Mu is the honor society for Industrial Engineers. UC Berkeley’s chapter of Alpha Pi Mu represents the Industrial Engineering and Operations Research department at UC Berkeley. Every semester we have a number of different events, from study groups to company information sessions to bonding activities like bowling and rock climbing. We work hard, but we still have fun in the process.
Katie Mak Event Organization APM offered a lot of chances for me to network and explore my professional career. I am interested in applying my skillset and knowledge acquired in IEOR classes to the healthcare industry to make the world a better place. I also love to travel and explore good food around the world. Here’s a picture of me in New York!
Wesley Jin Secretary The reason I chose to study IEOR is because I try to optimize all aspects of my life – whether it be finding the quickest route to class or getting the most for my money. As a second-year, I am still looking into the various career options that are open to me and being a member of APM allows me to do so with other motivated and high-achieving students. I love being outside, eating and cooking good food, playing basketball, and snowboarding!
Cesar Garza Recruitment Officer I want to make everything around me more efficient. For the future, I want to apply my Industrial Engineering and Operation’s research skills to pursue a career in real estate and facility management, making use of data analysis and optimization. My favorite thing about being an Alpha Pi Mu member is having people interested in the same subjects as me who are also exploring professional careers with me.
On October 7, 2015 the director of professional services at Navis, Pat Bukowski, came out to recruit IIE members for positions available in Optimization and Project Consultants. Navis is a global software market leader building and implementing advanced real time terminal operating systems and advanced optimization solutions for ports around the world.
From left to right, our social chair Waverly Runion, and our member Rebecca Martin, help pour drinks during our Welcome Pizza Social at La Val’s Pizza.
Institute of Industrial Engineers The Institute of Industrial Engineers (IIE) is the world’s largest organization supporting individuals involved with industrial engineering and operations research.
12 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
On September 10th, the Institute of Industrial Engineers hosted a Welcome Pizza Social, and invited all their members to come out and enjoy good company at La Val’s Pizza.
To the far left, our corporate chair, Jerry Cortez, facilitates questions at the end of the info session with Navis on October 7, 2015. Many students managed to hand in their resumes and were later called for interviews.
To see the latest news from IIE at UC Berkeley, visit: iie.berkeley.edu
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 13
IIE members Ebru Kasikaralar, Tinsley Zhu, and Jerry Cortez gather together and prepare go ice skating at Union Square in San Francisco for the Winter Social.
On December 6, 2015, the Institute of Industrial Engineers hosted a Winter Social at San Francisco Union Square. From left to right, members Katie Mak and Sherry Zhang, ice skate together and have a great time.
Members gather together on December 6, 2015 at Union Square in San Francisco. From left to right, Handi Tanuwijaya, Tinsley Zhu, Ebru Kasikaralar, Sophia Morgan, Theresia Wangke, Elisa Ramirez, Jerry Cortez, Carlos Medina, Sherry Zhang, and Katie Mak
14 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
PhD students Cheng Lyu, Kevin Li, and Quico Spaen are helping IEOR students connect in a new way by reviving and re-thinking an old program. The lunch and learn seminar brings together graduate students to discuss research topics, practice presentation skills, and learn about the latest software, tools, and techniques in the field. IEOR had a similar initiative in the past, but this time around the organizers are putting more focus on student needs to build the agenda. And with the IEOR department providing lunch to support the program, students can more easily participate. “We also had this tradition three years ago, but the format was that one student would present his/her work for the whole hour,” remarked Cheng, “Basically, he/she used it for just rehearsal purpose. So, if I went to a conference or was doing a defense I could use this forum to give a presentation for one hour. There are good and bad things [about doing it this way]. The bad thing is because our audience does not know much about the topic of each talk, that if they talk for an hour, it is not that effective. So, this time we want to make it more diverse, and look at what the demand is from the group — what they need — and then cater to that demand.” By focusing on what students need and working to deliver diverse content, the initiative is not only helping students practice their presentation skills, but serving them in other ways. For example, by allowing more than one topic per session, first-year students can learn about more topics in IEOR, helping them to make better decisions about what to research while here at Berkeley. The forum also gives students focusing on diverse topics in IEOR (e.g. energy systems, financial systems) a reason to connect with each other. Ultimately, since each student uses similar IEOR tools no matter the topic, these new connections can help give students a fresh perspective on their own research. The initiative is also focusing on getting students acquainted with the software and techniques that are becoming more important for the field. For example, the first session this year showed students Git, a version control software, and the basics of how to set up a Python development environment. “We see computation as being a big part of our research and our career. So, we had a session on Python — and not holding your hand through the entire process — but pointing you to the right resources, getting you started, and hopefully helping you get your hands dirty. Everyone is busy, so we want to make these things accessible in an efficient way,” said Kevin. For the next session, the group sought feedback and focused on a completely different topic. “At the second session we did a research and career panel because a lot of people were interested in doing internships, and wanted to know what other people in the department were working on. So, the idea of that session was to foster that interest and to get people knowledgeable about what other people are doing in terms of research, internships and jobs,” said Quico. Ultimately, besides helping students practice new skills, become acquainted with the latest tools, and learn about what their colleagues are researching, the lunch and learn initiative’s biggest value might be in helping students get connected in the first place. “This lunch and learn may be small but it’s already creating a better atmosphere in the department. When I first came here four or five years ago, first year students generally got together because of preliminary exams, but after the second year they disappear and you don’t know where they are,” said Cheng, who also mentioned that the current department renovations, which had a big focus on making the graduate offices more usable, also helped students connect this year, "Before the renovations only a few students were here working at school. I never saw others and didn’t know what they were doing or what they were up to. So, I felt isolated, especially when I got stuck on something. I feel that it is better to have someone to talk to besides just your advisor, because you can’t always see him or her. We think that by doing this seminar that we can get everyone more activated.” ■
Students Connect Through ‘Lunch And Learn’
The Lunch and Learn Seminar Group Back row: Cheng Lyu, Mo Zhou, Dean Grosbard, Salar Fattahi, Alfonso Andres Lobos Ruiz, Quico Spaen, Kevin Li, Tugce Gurek, Wei Qi, Sheng Liu Front row: Dimin Xu, Shiman Ding, Amber Richter, Xu Rao, Mengyu Li, Haoyang Cao, Yiwen Tang, Junyu Cao
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 15
This semester students came together for the first annual IEOR Graduate Student Symposium. PhD students presented on a broad array of topics representative of the many areas where IEOR tools are applicable including supply chains, energy systems, robotics, health care, neuro-imaging, and others. This first symposium was initiated and organized by graduate students with the goal of helping students connect and share research. “Yonatan and I were talking about how it would nice to have a one day event where all the PhD students could present their work,” said Auyon Siddiq, an IEOR PhD student who helped lead the initiative. “For the longest time, people would complain that research is very isolated and siloed,” said IEOR PhD student Yonatan Mintz who helped organize the symposium with Auyon. “It got to this point where it was like — we only have so many students in the department, but until today if you would have asked me what most of these people were researching I would have had literally no clue what they were doing at all.” So, Auyon and Yonatan decided to take action on their idea to have a one day conference. Besides wanting to help students learn more about what each other were working on, they also knew that students were looking for opportunities to practice their presentation skills. Plus, they thought it would be a good way to build community as it is difficult to get all the graduate students in one room as they are very busy with research, classes, and teaching other students. The only other time where students come together consistently is for the department seminar, but the speakers for this are chosen by the faculty. “We thought it would be fun to have something that was run by students,” explained Auyon. The faculty and the department also saw the value in the idea and gave their full support which Auyon and Yonatan said made things much easier. The department provided food for the event and helped book the room, but most of the heavy lifting was done by the students themselves to organize twelve speakers who each gave a twenty minute talk. Speakers, presentation titles, and photos for IEOR Graduate Symposium 2016 can be found below. 1. Clay Campaigne - Comparing Welfare Effects of Retail Electricity Tariffs by Simulation 2. Andrés Gómez - Simplex-Based Methods for Programs with Conic Quadratic Objective 3. Cheng Lyu - Improved Algorithms Solving Generalizations of Isotonic Median Regression 4. Auyon Siddiq - Inverse Optimization with Noisy Data 5. Animesh Garg - Semi-Supervised Learning in Robotics 6. Quico Spaen - Novel Algorithms for Automatic Segmentation and Signal Extraction in Calcium Imaging Data 7. Yonatan Mintz - Scheduling Personalized Weight Loss Interventions Using Predictive Analytics. 8. Mo Zhou - M-CAFE V1.0: Motivating and Prioritizing Ongoing Student Feedback Using Collaborative Filtering 9. Morteza Ashraphijuo - A Strong Semidefinite Programming Relaxation of the Unit Commitment Problem 10. Salar Fattahi - Design of Optimal Static Distributed Control 11. Jiaying Shi - Stochastic Unit Commitment with Topology Control Recourse for Renewables Integration Shiman Ding (not pictured) - Centralized and Decentralized Supply Chain Collaboration 1
1
2
16 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
Students Organize First Annual IEOR Graduate Student Symposium
9
6
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 17
4
10
7
5
8
11
3
Project Title: NetApp Predictive Modeling of the Sales Pipeline Project Description: The project is to provide Netapp, a storage-industry based company, a precise prediction on the expected sales amount in a given time frame. The team makes use of machine learning and predictive modeling techniques to model dynamics contained within the conversion between the given prospective deals and deals that are successfully closed. What is the project for? NetApp Inc. What is the potential impact of your product? Though the model is primary developed for in-house sales prediction, the mechanism is believed to be applicable for most companies who have a sales pipeline. The prediction will benefit the company by providing other business units actionable recommendation, including behavior change in sales teams and better budgeting in the finance team. What is the most valuable thing you’ve learned this year? We not only have learned various techniques in building models, we also have acquired the crucial skills in participating in industry projects. We truly understand the importance of communication with clients/advisors and collaboration within the team. Sometimes the team might not be equipped with all the skill-sets that are required, but with good collaboration and a willingness to help and learn, things work out.
Master of Engineering Project Profiles Master of Engineering (MEng) is an applied degree combining business-oriented coursework with applications-focused industrial engineering and operations research courses. Below we profile two great projects. this year.
Team NetApp — Back row (left to right): Camille Lehujeur, Kar Fu Kong, Akshee Jain. Front row (left to right): Karl Sahyoun, Joel Varghese
18 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
Project Title: Forecasting Retail Deposits using Macroeconomic Drivers Project Description: Project first goal is to identify and isolate the most important macroeconomic variables which drive (affect) bank deposits (certificates, checking and saving account) in the USA. The next goal is to build predictive models to forecast these deposits and their change on a monthly time frame using these macroeconomic variables. The final goal of the project is to derive any other insights from deposit data to possibly better understand consumer behavior. The product therefore is a statistical model and its implementation in R. Who is the target market for your product? Project is meant for Nomis Solutions- our industry advisor and a company making price optimization software for commercial banks. The ultimate aim for Nomis is to include our model into their software which will be delivered to their client banks. Some other industries like private equity firms might also be interested in such a product. What is the potential impact of your product? Nomis manages about $1 trillion from its clients out of which deposits forms the major portion (~ $400 billion). Nomis’ product due to its predictive nature can help profit improvement by up to 18% (all data based on Nomis official website). So although it is difficult to exactly quantify the final impact in monetary terms, it would be safe to say that our model can help save $1 billion if not more — so our project is a billion dollar venture. What is the most valuable thing you’ve learned this year? Working in and managing cross-cultural teams.
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 19
Team Forecasting Retail Deposits— (left to right) Mihir Tamhankar, Xueqi Wang, Arthur Cahen, Yin Bian (not pictured)
Master of Engineering Project Profiles (continued)
Introducing the 2016 Sutardja Center #WHATSNEXT Watchlist
Berkeley team wins geisinger health collider
Sutardja Center
The Sutardja Center is pleased to announce that Team Outliers (Rebecca Barter and Shamindra Shrotriya) have won the Geisinger Health Collider. Team Outliers competed against four other teams: three teams from Goethe University in Germany and one from Stanford. Rebecca is a PhD student in the Department of Statistics at UC Berkeley (advised by Professor Bin Yu and Professor Jasjeet Sekhon), and Shamindra is an MA student also in the Department of Statistics. For their project, Team Outliers focused on an analysis of Chronic Obstructive Pulmonary Disease (COPD) and its association with environmental factors. Students in the competition were given access to aggregate electronic medical record data to help Geisinger predict personalized treatment outcomes, better allocate healthcare resources, and obtain early warnings in scenarios of crisis. As the winners of the competition, Team Outliers will receive paid internships from Geisinger Health. “This project provided us with a rare and unique opportunity to get some hands-on-experience dealing with real electronic medical record (EMR) data from Geisinger,” said Rebecca Barter from Team Outliers, “There are so many questions that one can ask of such a rich, yet typically underutilized source of data, the answers to which truly have the potential to make a difference in the field of medicine. Further, the problem of blending this EMR data from Geisinger with external environmental data presented a unique set of challenges which, together with various statistical and machine learning methods, allowed us to tackle an even broader class of questions, such as how does the weather affect pneumonia? or does the level of pollution in a neighborhood influence ones risk of developing COPD?” Rebecca and Shamindra were interested in discovering factors that help predict whether a new patient who has been diagnosed with pneumonia might also have COPD. To achieve this goal they combined data from electronic medical records (provided by Geisinger) with publicly available data on pollution and the weather. Ultimately, they were able to construct a model to help predict undiagnosed COPD for new pneumonia patients, and also make suggestions about the types of data that Geisinger could track in the future to improve the model’s accuracy. “Our team was delighted to read such a clean and well researched investigation,” said Dr. Nicholas Marko, Chief Data Officer at Geisinger Health of Team Outlier’s project, “Their figures were impeccable (and more importantly, relevant), and their introduction / risk factors sections demonstrated that they really did their homework on the subject.” ■ Download Rebecca and Shamindra's winning presentation here.
Team Outliers: Shamindra Shrotriya and Rebecca Barter
IEOR's Sutardja Center for Entrepreneurship & Technology has published its What’s Next Watchlist for 2016. The Watchlist was created by tracking applied research and venture development and aims to determine which emerging areas are the best investment bets for 2016 based on timing and potential impact. What’s Next Watchlist Emerging Areas to Watch for 2016: 1. Healthcare + Data / Cognitive Technology 2. Cybersecurity and Encryption 3. Data+Cognitive Technology+Enterprise 4. Drones and Robotics 5. Connected Cars 6. 3D Printing, Digital Manufacturing, and Verification 7. Fintech, Transactions, and Blockchain 8. Education and Training + IT/Data 9. Collaboration, Workspace, possibly with Virtual Reality 10. Media+Brands+Retail and Data/Cognitive Technology 2016 Pre-investment category: 1. Food Sustainability and Meat Alternatives 2. Neurological Disease Prevention Correlated Additions from Advisors/Investors: 1. Broadband and 5G Infrastructure Methodology: This list was obtained by tracking two sources of data. One is the technical reports of UC Berkeley’s Sutardja Center projects. Authors represent Silicon Valley Executives, Ph.D. students, and undergraduates. The second source is from the Series A investments over the past 12 months from leading VC firms that include Andreessen Horowitz, Kleiner Perkins, Sequoia Capital, Foundation Capital, Greylock Partners, Khosla Ventures, and others. Data from the two sources was combined to create one list. The list was then voted on for timing for investment in a range from “too early” to “crowded” by investors from the Sutardja Center’s Board and an extensive network of investors and industry leaders. Data collected was analyzed using data science tools for variation errors and ranked according to the strength of the data signal for each potential area. PI: Ikhlaq Sidhu. Project team: Carlos Rodrigo Ramos De La Vega, Jocelyn Weber Phipps ■
20 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
Blockchain Collider Challenge: Come up with the best possible use case for blockchain technology. Winner: Team All in Favor (Prize: $5k) Imprint Energy Collider Challenge: Create new battery specs to help product designers understand how to use Imprint Energy’s flexible printed battery tech. Winners: Team SymbIoT ($4k) & Team Battery-Tech ($4k) Geisinger Health Collider Challenge: Use Geisinger’s massive electronic medical record data to create personalized treatments, better allocate healthcare resources, and obtain early warnings in scenarios of crisis. Winner: Team Outliers, Prize: Paid internship with Geisinger Big Data in Retail Collider Challenge: Help retail firms understand how to use big data to increase profit: what data is available, what skills are needed, and what tools are most valuable. Winners: Team $100k level ($4k) & Team $1M Level ($1k) Intellectual Property Collider Challenge: Create go-to-market strategies for Berkeley intellectual property to assess the size of the market and opportunity. Winners: Team NOG ($5k) WaterSeer Collider Challenge: Develop a prototype and use case for WaterSeer patented technology which creates water from air in arid environments. Winners: Team Stratus ($6k) & Team HydroLily ($5k)
Sutardja Center Collider Winners for Spring 2016 What is the innovation collider? The Sutardja Center helps create an innovative environment by bringing together diverse people in academia and industry. The Sutardja Center then uses experiential learning to teach individuals how to have an innovation mindset that is reflective of the most successful entrepreneurs. These diverse individuals with the right mindset then bring their unique skills together to form teams which result in innovative companies and ideas. One way the Sutardja Center challenges students to create innovation is through collider projects. These projects ask students to form teams to solve a problem for an industry partner. Below you will find a short summary of each project and the winners for 2016.
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 21
22 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
Be honest with yourself and others (i.e. don’t just say what you think others want to hear). Work hard on a problem to reveal its secrets. Failure is an essential part of the innovation process — but make only new mistakes. This is the advice that businessman and investor Ben Horowitz shared with students in a fireside chat with Professor Mike Franklin, the Thomas M. Siebel Chair of Computer Science and director of the AMPlab at UC Berkeley. Horowitz was interviewed as part of the A. Richard Newton Series which brings distinguished innovators from around the world to Berkeley to share their successes and failures with students. Horowitz brought a unique perspective as co-founder of Andreessen Horowitz, one of the valley’s top venture capital firms with a portfolio of investments that includes Facebook, Twitter, 500px, Foursquare, Buzzfeed, Groupon, Lyft, Oculus VR, and many more. Starting with an initial investment of $300 million in 2009, Horowitz and partner Marc Andreessen’s firm is currently valued at an estimated $4 billion. Honesty So how did Horowitz become one of the most successful venture capitalists in history? Horowitz, a Berkeley native (and Berkeley High School alum), left the Bay Area to attend university at Columbia where he was shocked by how different New York was. “You really want to fit in, particularly if you go to a college like I did where I didn’t have any friends and was from a weird hippie town. So, I show up there, and I really want to make friends, but you have to be careful because your friends in college are… kind of stupid.” Horowitz’s friends helped him learn his first life lesson: to trust himself and not always try to please others. After attending a computer science class, he was inspired by the English mathematician Alan Turing and his idea of ‘Turing completeness’ — the basic idea that a general purpose computing machine could simulate all kinds of machines without having to build a different machine for every single purpose — an idea that paved the way for how we use computers today, “Even though Alan Turing was 40 years before then, I was reading about this guy going, Wow, this guy has a secret that is going to happen to the world that nobody knows about but me. That’s awesome, I’m going to major in computer science.” But at the time, computer science wasn’t as respected as it is today, so when Horowitz told his friends that he was planning to study CS, they teased him. Inspired by Turing’s secret, he knew he needed to trust himself, “Even though I really wanted them to like me because they were my friends, I wasn’t going to listen to them. That little lesson was the key to everything else that I think happened to me in my life, because every time you have a breakthrough idea, a really innovative idea, by definition it’s going to look like a stupid idea. If it looked like a good idea, it wouldn’t be innovative, it would be something that everyone else already knew and would be obvious.” So, in spite of his friends, Horowitz studied CS at Columbia and eventually received a Master’s degree from UCLA where he worked under Professor Leonard Kleinrock, who is credited as a pioneer in the invention of the internet for his contribution of packet-switching. After UCLA, Horowitz worked for the Lotus Development Corporation helping to develop LotusNotes, until he heard of something called Mosaic. “What LotusNotes was doing in the most complex way imaginable, Mosaic was doing in a completely open and simple way, and I thought — this is the future, and I’ve got to go. So, I applied at a company called Netscape.” Again, Horowitz’s ability to act on seeing the next big secret was uncanny. Mosaic, the first modern web browser, is credited by many for popularizing the World Wide Web. It later changed its name to Netscape and launched Netscape Navigator. (Incidentally, Mosiac’s code was licensed to Microsoft who used it to create the first Internet Explorer — which started the so-called ‘browser-wars’ of the 1990’s.) While Horowitz was undoubtedly shrewd in his ability to recognize the potential of the browser, what he (may not) have anticipated when going to Netscape was that there he would meet his future partner, Marc Andreessen. Andreessen (who was just twenty-four at the time) was co-author of the Mosaic browser and co-founder of Netscape. During the talk, Horowitz lauded Andreessen’s intelligence, “I swear to this day that the reason that Marc invented the browser was that in those days, there just wasn’t enough for him to read.” Expanding on his earlier lesson to trust himself, Horowitz shared what he believes to be one of the most important skills anyone can have — honesty. “[One] thing that is very hard is honesty. That means honesty with other people and honesty with yourself. The truth is, there is not a naturally honest person in the world. Nobody really is, it’s something you have to work at. And the reason you have to work at it goes back to that you really want to feel good about yourself, and you want people to like you. And the way to get people to like you is to tell them what they want to hear — not the truth — you tell them what they want to hear. That’s one of the biggest challenges if you want to lead an organization or do something important. You can’t lie to other people or lie to yourself. If you do, they will see through it. You might think you are being slick — but they see through it.” “Telling the truth is work. The next time you say the thing that makes you feel better — you need to check yourself and try to be honest. If you learn to do that, that’s actually an amazing skill.” Failure Horowitz went on to share some of the mistakes that he has made in the past. He discussed an early $30 million mistake when he expanded the office space of he and Andreessen’s early company Loudcloud, one of the first companies to talk about cloud-computing and software as a service, which later became Opsware (a company sold to HP for $1.6 billion in 2006). Horowitz had expanded the space in 1999 right before the dotcom crash which reduced his $30 million investment tenfold. “If you try to do something really hard, you’re going to make mistakes. Entrepreneurship is very much like scientific discovery. If you read about many of the great scientific discoveries, they almost never found what they were looking for… you rarely end up doing what you set out to do. By definition, the whole endeavor is a big mistake.” Horowitz elucidated by telling the story of Stewart Butterfield, the entrepreneur who created Flickr, the popular photo sharing app. Flickr was not conceived of as a photo-sharing application but was originally part of a massive online multi-player game called Game Neverending (the Flickr part was a photo uploader for the game’s avatar). Later, in 2009, Horowitz invested in Butterfield’s new game called Glitch. Unfortunately, the game was created with Flash (a technology that Steve Jobs announced in 2010 that Apple would not support on its extremely popular iOS devices which were already transforming the way games were being played) and did not catch on. So, when the game looked as if it would fail, Horowitz received a call from Butterfield, “Ben, I’ve got $6 million left. I have no way to raise money to solve the issues, and it will cost more than $6 million to finish the game. So, I’ve got like three choices: pray for rain and try to finish, I could shut down the company and give you your $6 million back — or we built this tool that we use in our engineering team to communicate with each other and make engineering work a little better, and we could put that out as a product.” While skeptical of completely pivoting Butterfield’s venture — from a video game to enterprise software — Horowitz told Butterfield to go for it. And it paid off. The team messaging app that Butterfield developed to help his engineering team communicate was called Slack, a company that in 2015 was valued at $2.76 billion — and an endeavor that started as a total mistake. Secrets “A lot of the process is, you have a secret, something that no one else knows. Because [Butterfield] was like pulling his hair out, literally smoking eight packs of cigarettes a day, he was doing everything he could do optimize development. So, he learned where software development was sub-optimal. He discovered a secret in the process of trying to build a game, and it was that secret that led to the company. And that is a really big key. You have to go do something really hard if you want to learn something about the world. And it’s not that nobody else knows it — but almost nobody else knows it. Nobody who does know it, is acting on it. And that’s when you have a breakthrough. That’s when you have something you can build a company around, that’s when you have something that you can build something really important around. But it starts with hard work.” Many lessons can be drawn from Ben’s unconventional approach to investing. As someone who has led Silicon Valley’s greatest startups and influenced the impact computing and media has had on our lives, Ben’s vision for the future, and his approach to investing serve as an example to all entrepreneurs striving to make it in the tech world. His greatest asset though, remains his passion for hard work and honesty. ‘Stay strong in the struggle,’ Ben wrote on everyone’s books, as he passed on wisdom and inspired students after his talk. ■
Ben Horowitz Talks Honesty, Failure, and Secrets at Newton lecture series
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 23
On Thursday, March 10th, alumni of the Sutardja Center’s Engineering Leadership Professional Progrom (ELPP) gathered at Samsung’s newly opened North American HQ in San Jose for a well-attended mixer. The event included tours of the “vertical garden tower”, a $300 million, 1.1 million-square-foot research and development center, a highlight of the evening. Another highlight was an update by Prof. Sidhu on the Sutardja Center’s activities and their relation to topics and analyses done in ELPP, such as the Center’s recent Blockchain Collider that leveraged work done in ELPP as a foundation for instruction of Berkeley students on this important, new fundamental technology. An additional highlight was the attendance by Pantas and Ting Sutardja, who were able to speak with alumni of the program, as well as enjoy a tour. To view Ikhlaq Sidhu’s slides from that evening, see ELPP-Alumni-Spring 2016-Shared. ■
Engineering Leadership Alumni Gather at Samsung’s New HQ for Mixer
24 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
For more information about the Sutardja Center’s Executive Leadership Professional Program visit: scet.berkeley.edu/elpp
Ricardo San Martin (2nd from left) and Magdalena Hevia (far right) meeting with innovators at SF Maker Faire
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 25
Building a New Bridge to Silicon Valley Culture
The Sutardja Center's Global Partner Program brings scholars from around the world to Berkeley to learn about innovation and entrepreneurship. To learn more visit: scet.berkeley.edu/ global-partners-at-uc-berkeley
By Ricardo San Martin, Visiting Professor & Magdalena Hevia, Director Bridge Executive Program, Pontificia Universidad Católica de Chile Nowadays many companies from all around the world come to Silicon Valley (SV) to learn how to become more innovative and technologically competitive. However, the recurrent question is: what’s the best way to do that? At the School of Engineering of Pontificia Universidad Católica de Chile (PUC), we are testing a new program called Bridge Executive to help Silicon Valley visitors learn about innovation more effectively. By doing so, we are leveraging our partnership with the Sutardja Center for Entrepreneurship and Technology at UC Berkeley and its collider program. Our conviction is that the traditional approach, where executives come to SV to purchase technologies is not enough. We believe the most effective way to learn about SV innovation is to experience its ecosystem, learn its frameworks and understand its mindset to be able to come up with innovative ideas and set up teams to execute them. This approach normally encounters major difficulties, such as lack of time from top executives to really understand how ideas are executed in SV, and a very traditional internal culture that resists change. However, most companies realize today that there are no other options: they either have to rapidly adopt innovation frameworks or be displaced by novel and fast growing technological companies. Bridge Executive provides a platform where Chilean companies can rapidly enter into Silicon Valley networks through our connection with the Sutardja Center, work on pressing technological problems and gain insights about emerging technologies by working with PUC engineering students who stay 4-5 weeks in SV. Moreover, they simultaneously start addressing internal cultural issues that may be inhibiting innovation. The prototype program that started in January 2015 considers the following activities: Selecting leading Chilean companies that are aware that the new technologies created in SV will certainly disrupt their business models. Choosing a particular project/problem they want to develop with Silicon Valley. Bringing top company executives to SV for about a week. Prior to coming, these executives are trained remotely on the mindset and cultural aspects of SV by the PUC team in SV. Select 1-2 senior students per company that stay for at least 4-5 weeks in SV to understand if the problem is correctly framed, if the timing is right and which partners and other stakeholders might help in the execution of the project. These students benefit from the networks provided by the Collider, and therefore have a vantage entry point to SV. The role of the students is not only to try to solve a specific challenge, but also to provide new ideas and emerging technologies to the company. Evaluation of the innovation culture of the company using the recently developed Berkeley Innovation Index to assess what are the main roadblocks to execute innovative projects. These topics are addressed through conversations with the executives of the company. The emphasis is placed on team dynamics and on the conversations that have to take place to really open a new space of opportunities for the company. Follow-up conversation between the PUC team in SV and the company’s team to facilitate the execution of the projects, and move towards a more innovative culture. We expect that the program will help Chilean companies adopt current SV innovation practices that will help them maintain their competitiveness in the region. ■
This image from Matrix, a UC Berkeley newseltter, shows Shmuel in 1983 shortly after joining the Berkeley faculty
26 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
"Ultimately, that’s what these makes this whole area so fascinating for operations research and industrial engineering — that electricity is probably the most complicated supply chain in the world."
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 27
“You cannot expect to run system based on altruistic behavior, which is like trying to design an airplane based on levitation. Greed and profit-seeking is like gravity — it’s there and cannot be assumed away.”
28 SPRING / SUMMER ‘16 | IEOR Alumni Newsletter
IEOR Alumni Newsletter | SPRING / SUMMER ‘16 29
B.A. Operations Research & Management Science Michele Gleit Anirudh Srinivas Kilambi Andrea Kwan Jong Yoon Lee Kevin Chia-Hau Lee Tiffany Soebijantoro Cong Wang Cynthia Wu Linnan Zheng B.S. Industrial Engineering & Operations Research Spring 2016 Thomas P Boeck Alejandro Castillejo Cheuk Him Desmond Chan Albert G Chandra Burt J Chen Courtney Zhi-Jia Chow Ryan Clifford Corley Catherine C. Darmawan Yirong Fan Oliviero Felice F. Figus Cheyanne Isha Galinato Calvin Gozali Li Gu Chunchun Huang Farzad Kargaran Rajit Kinra Sooyoun Kwon Jinzhou Li Nanavati Mei Yi Low Hing-Man Mak Riley John Murray Bowen Ni Sasha Mehul Patel Dylan Labatt Randle Raymond Rudianto Jerome T. Rufin Kunal S. Shalia Jonathan Sumarto Sirisha P. Varigonda Jessica A. Wijaya Hiromichi Yamamoto Yikun Yang Master of Engineering Spring 2016 Marie-Camille Achard Utsav Akhoury Ana Sofia Aramayo Carrasco Xabier Azagirre Omar Bentahar Yin Bian Yacine Brini Arthur Laurent Cahen Guillaume Candeli Anis Chagar Baihan Chen Rebecca Chew Min Ting Kevin John Danser Marie Douriez Boran Fu Martin Gouy Xiaoyu Gu Shaoting Han Ziang Hu Akshee Jain Daqian Jiang Jason Kam Hyeji Kim Kar Fu Kong Camille Lehujeur Juanyan Li Lujia Li Yifang Liu Zhenhang Liu Tian Luan Thibault Martin Armelle Patault Nihit Prakash Neng Qiao Honghao Qiu Elana Rosen Karl Paul Sahyoun Kilian Schindler Lydia Marie Smith Nick# Stauss Xilin Sun Shishir Sura Mihir Subodh Tamhankar Chong Wee Tan Ludovic Thea Xuedong Tian Cedric Roch Vallee Joel Prince Varghese Chaitanya Pradeep Wadi Tingting Wang Xueqi Wang Sio Fong Wong Yiteng Xu Xuanqi Yang Yiheng Yang Bing Lin Yeo Wendi Zhao Zirui Zhong Chuanhai Zhu Master of Science Spring 2016 Huseyn Abdulla Aaron Warner Atton Agrima Bahl Deepa Dilip Kalelkar Mengyu Li Nishant Yash Lokhandwala Shir Nehama Aurelien Amiyao Ouattara Yiwen Tang Kevin Maxwell Vale Ji Wang Haoyun Wu Xijia Wu Qi Xu Dimin Xu Ph.D. Fall 2015 Dan Bu Paula Lipka Wei Qi Spring 2016 Qiaochu He Amber Richter Peng Yi Qi Zhang
Our Graduates
FALL 15 | IEOR Alumni Newsletter
Alumni re-uniting at the spring social in March
IEOR renovations continued with a new state-of-the-art lobby to welcome visitors
Graduation Day!
Stay connected Website: ieor.berkeley.edu LinkedIn alumni group : https://www.linkedin.com/groups/1781 Facebook: www.facebook.com/BerkeleyIEOR Donate to IEOR Help us continue to deliver world-class research and academic programs • Give online by visiting give.berkeley.edu
2016 IEOR Award recipients Undergraduate Faculty Award - Riley Murray Outstanding ORMS Student - Linnan Zheng Alpha Pi Mu Award - Aaron Prohofsky - President Bowen Ni - Vice President Mehran Navabi - Treasurer Wesley Jin - Secretary Cesar Garza - Recruitment Officer Katie Mak - Event Organization Institute of Industrial Engineers Award - Yirong Fan - President Elisa Ramirez - Vice President Albert Chandra - Finance Chair Sherry Zhang - Technology Chair Waverly Runion - Social Chair Jengyee Prize - Srushti Vora Departmental Citation - Jessica Wijaya MEng Intellectual Contribution Award - Chong Wee Tan Katta G Murty Prize - Morteza Ashraphijuo Auyon Siddiq Outstanding GSI Awards Carlos Deck Kevin Li Rebecca Sarto Basso Quico Spaen IEOR First Year Faculty Fellowship Award- Arman Jabbari IEOR Faculty Fellowship Award - Quico Spaen
Professor Kaminsky awarding the Jengyee Prize to Srushti Vora
Volunteers for Cal Day: Sarina Chen, Taylor Lyberger, Theresia Wangke, Stella Bao, and Vincent Liu
Dean Sastry joins the IEOR faculty in celebrating Shmuel Oren’s election to the National Academy of Engineering.