When I was building this website, I read many online biographies of peers and people that I respect in my industry. I appreciated that they were mostly factual, brief, and humble. But I always walked away wishing I could understand more about how they ended up where they were and why they made the choices they did. For that reason, I’ve attempted to provide a somewhat longer-form bio with a bit more raw information. This is ordered from most to least recent.

The information I try to convey with each job/career step is:

  1. Hypothesis for the change
  2. How I accomplished the change
  3. What I learned from the experience

Sabbatical & advising (2022 – present)

Currently, I’m taking time off from full-time work to recharge and figure out what’s next for me. In the meantime, I’ve founded Teraton to try to help early climate startups go “from 0 to 1” and to help more established companies improve their data science approach. If you think Teraton could be helpful to you or your impact venture, you can reach out here.

While I’m not certain exactly what I’ll do next, the new things I’m looking to do are:

  • Working on products which have more quantifiable impact (e.g. kg CO2eq)
  • Diversifying the impact of my “9-5” by working across multiple companies (instead of investing all of my effort into another single early-stage startup)
  • Learning new skills other than individual technical work (e.g. management, strategic advising, investment)
  • Avoiding burnout with a flexible and reduced work schedule (e.g. 4-day weeks, 80% roles, etc.)

Data scientist, Myst AI (2019 – 2022)

I started working at Myst AI in 2019. My hypothesis for joining Myst was to get back to the original reason I left Genability: I thought I could be more fulfilled and have more impact if I worked at the intersection of math and CS on an important problem (as opposed to software engineering alone). I got the position because friends of mine referred me when Myst was looking to make their first data science hire and I had enough energy industry, startup, and software engineering experience to justify taking a risk on my lack of data science experience.

Myst turned out to be exactly what I was looking for, enabling me to develop time series forecasting models, build cloud infrastructure for experimentation, and improve research efficiency via automated machine learning. I’m also grateful to have learned from brilliant colleagues who had better technical vision (Pieter Verhoeven), business strategy (Titiaan Palazzi), and data science chops (Erin Boyle) than I.

I left Myst in August 2022 to recharge, pursue a new type of role, and search for ways in which I can have a more quantifiable impact.

Co-founder and CEO, Nanogrid (2017 – 2019)

Prior to Myst AI, I was the co-founder and CEO of Nanogrid. Nanogrid originated at Powerhouse’s SunCode Hackathon. In starting Nanogrid, I hypothesized that I could learn about starting a company, do more data science-heavy work, and have high-leverage impact by supporting the growth of distributed energy storage.

As with most startups, things did not go as planned. Ultimately, I did not do much data science work and instead learned a lot about starting a company and negotiating contracts for a technical product. I also grew my network substantially, which led to the opportunity to be the first hire at Myst AI.

In terms of learnings, I became intimately aware of the:

  • Importance of sales cycles (e.g. large utilities can take 6 months to 2 years to make a product decision)
  • Need to balance building vs deciding what to build (e.g. decide on objectives for a 3 month period and commit)
  • Value of warm introductions and industry connections
  • Difficulty of building something actually useful
  • Importance of business strategy and technical execution (vs the initial idea)

I also learned a bit about my own theory of change. Initially I had thought I would be okay pursuing any role and any product as long as I thought it could have a high scale impact. After experiencing pretty extreme burnout I developed an understanding that, for me, I need to be doing work that I think I’m well suited towards and that, in the long run, I’ll have the most impact by developing expertise in those key areas (overlap of software, math, and connecting with people).

I’m incredibly grateful to the folks who helped make Nanogrid happen and proud of my co-founders Dan Lopuch and Jon McKay for leading the company to a successful acquisition (by Arcadia).

Software engineer, Genability (2015 – 2017)

Initially I thought a CS person looking to work in clean energy was rare. But after googling “software and renewable energy”, I discovered the bay area Software for Renewable Energy meetup group. This group is how I became aware of the Oakland-based accelerator/investor, Powerhouse. After attending their annual clean energy party, New Dawn, I emailed every sponsor and every person I had met. Out of that networking sprint, I landed a job at Genability which provides electricity rate data and cost modeling for almost every clean energy company, from solar providers like Sunrun and SunPower to electric vehicle companies like BMW, Tesla and more.

At Genability I learned many things about modern software development (agile, CI, CD), cloud-based infrastructure (AWS), and the good/bad of startups. After a few years, I realized that while learning to be a full-stack software engineer was useful, I thought I could be more effective (and more fulfilled) if I also applied my background in math. With little research/thought, I decided on energy storage for my next move. My general reasoning was that the cost of solar was already below $1/watt and that marginal effort spent on energy storage would perhaps provide larger marginal emission reductions.

Graduate school, CMU

At the end of my undergraduate education I had a diverse set of mostly orthogonal activities. Instead of looking to continue them independently, I looked for ways to merge my interests in public service, climate change, and computer science. After limited searching, a friend recommended PhD programs which combined various forms of engineering and public policy. I applied to many of these programs, my first choice being Engineering and Public Policy (EPP) at Carnegie Mellon University (CMU). Having not done any research as an undergraduate or produced anything similar to the work done in those programs, I was rejected by all of them. By a stroke of luck, CMU offered me a spot in their Civil and Environmental Engineering Master’s program.

I took the opportunity at CMU and spent a year in Pittsburgh, PA. This allowed me to take courses in EPP, learn a bit more about power systems, and continue my analytics education in machine learning / optimization. Thanks to some direction from my advisor, Jared Cohon, I quickly discovered that I enjoyed optimization and had no business as a Civil Engineer. I also made an observation about EPP: while applying rigorous quantitative methods, EPP primarily studied existing solutions to important problems. I felt that before I could participate in evaluating technological solutions, I should try creating my own. For this reason I applied to another set of PhD programs, this time in the field of Computational Sustainability. Again I was rejected by every program I applied to (including CMU). Partly because of a poor application strategy, partly because again I had done no research at CMU and had published no papers.

Discouraged, I decided I would try to work as a software engineer or data scientist in the clean energy industry.

Undergrad, UC Davis

From my perspective, UC Davis would allow me to continue the life I had lived in high school, split evenly between school, sports, and some form of academic extracurricular. I was a political science major, but also took classes in math. My sport was throwing javelin. After a chance encounter, my extracurricular activity became leading a council group for the campus dorms.

The first thing to change that perspective was being cut from the track & field team. At the same time, I had begun an internship in student government. In a scramble to re-align my self identity, I decided that I shouldn’t spend 4 hours every day training anyway and that public service was where I should focus my attention.

The second shift was academic. After deciding I might double-major in math and political science, I enrolled in my first computer science (CS) course taught by Vladimir Filkov. With some help from my dad, hundreds of hours, and a great deal of guidance from Professor Filkov, I made it through the course and fell in love with the field.

From then on I slowly transitioned away from political science and towards math and CS. My feeling at the time was that in political science I was studying the problems I thought were important, but wasn’t learning the tools I could use to solve them. While I viewed problem discovery as important, I concluded that I should focus on developing math/cs skills and learn real-world problem-solving through student government instead. I was able to make the switch to CS thanks to the flexibility of the CS major (at that time) and some salient advice from Cal Newport via his college-focused blog, Study Hacks – highly recommended. While counter-intuitive to me at the time, the Study Hacks idea of under-scheduling proved the most useful. For example, I originally intended to finish my major in political science but instead, I gave up on even minoring and dramatically reduced my courseload to focus on a smaller set of CS and math classes. This gave me much more time than my peers to work on homework assignments and ultimately allowed me to take advantage of other low-hanging fruit.

While I was having my CS “renaissance”, I also increased my involvement in student government. I was elected to their “senate” as one of twelve senators responsible for setting a $12M operating budget. Key projects of mine included “bait bikes”, a program to reduce bike theft, and an effort to enable students to teach courses championed by my friend Rylan Schaeffer. Similar to my previous experience with Speech and Debate, student government allowed me to hone my public speaking. It also taught me a great deal about how decisions in any organization get made. Initially I thought that having the most concise, well-reasoned arguments would magically win over my peers and garner votes for my legislation. Instead I found that befriending my peers, establishing and aligning values, and compromising on various issues was much more effective than argumentation. Looking back, my biggest failures were a lack of follow-through and an inability to mentor and delegate. I did not monitor or measure the performance of “bait bikes” and I did not pass down my learnings to other interested UC Davis students so that they too could participate in student government.

Internships

During college I had two summer internships. One was at the US Department of Education, the other was at Apple.

At the Dept of Ed. I was energized by the problem but frustrated by the tools used to solve them (mostly Excel and a slow interface to a SQL database). I did some useful work, but mostly assessed data quality. I also ended up taking their servers down for a few days (on accident) but that’s a story for in-person.

At Apple, my work on Binary Classification was more challenging, which I enjoyed. But the problem itself seemed to not be of any critical importance. Who cares if people with iPhones can auto-magically make their photos look a little bit better?

Based on these two experiences, I resolved that I would try to do technically challenging work on problems I thought were important.

Pre-college

Before college, my “professional” time was split between school, sports (rowing and discus), and debate. While I earned various awards in all three, I excelled at none. I was fairly sure I was not going to get into a “top” school and my parents (thankfully) encouraged me to only apply to public universities.

I was rejected by UCLA and UC Berkeley and narrowed my choice to be between UC Davis and the University of Washington. I chose UC Davis because I was accepted into the honors program and because that was where my partner at the time wanted to go. At Davis I could also be a walk-on for the track & field team.