Early Life

I grew up in Half Moon Bay, a coastal town shrouded in fog. Not quite the sunny beach paradise my parents desired. My dad’s from Portugal, so we visited every summer–that was fun. In elementary school we moved to Portugal for family business, which was less fun. I coped with the social isolation by burying myself in fantasy books. That obsession is with me to this day.

Three years later we returned to HMB where I attended middle school and high school. I was a sporty nerd, getting good grades while playing basketball, tennis, and water polo. I miss the camaraderie of team sports immensely. I avoided alcohol and parties until the very end of senior year, while indulging in video games and LAN parties–cementing my status as a geek. I learned to program through Minecraft plugins, and soon modding the game was more fun than playing it.

I also started HMB High’s robotics club, which likely secured my admission to UC Berkeley where I decided I would professionalize my joy of writing code. I was torn between pursuing physics and computer science, but hoped that robotics would split the difference.

College

To over-correct for my innocence in high school, I immediately joined a fraternity at Berkeley. Perhaps unsurprisingly, there were more EECS majors–computer nerds like myself–in my fraternity than in any other. Classes were hard, but every class had at least one friend to commiserate with. I vividly recall the shock of my first 40% exam score, thankfully I was not alone. While I suspect that my participation cut my GPA a few points, I firmly believe the friendships outweighed any cost.

I avoided humanities like the plague and overcommitted to grueling technical classes that I found fascinating. Nonetheless, Berkeley forced us to take a few writing and “social justice” classes to broaden our horizons. Like a fish unaware of water, it is only in hindsight that I realize how ideologically driven those classes were. I’m glad that the engineering classes were mostly isolated from such nonsense, but I was not unscathed. I was vegan for a half dozen years and educated others on the environmental impact and ethical superiority of avoiding animal products. I no longer restrict my diet, but do still feel sorry for victims of factory farming. I suspect that the real solution requires more knowledge and growth, not draconian restrictions and anti-growth mindsets.

Tesla

In keeping with my technical proclivity and sustainable priorities, I was overjoyed to receive an internship my junior year and then a full time position at Tesla when I graduated. I thought after four years of rigorous study that I was an expert coder, but that assumption was quickly corrected. I probably learned more in the first two years at Tesla than that previous 4 years at Berkeley.

I worked on the Fleet Analytics team, designing and implementing data pipelines to process hundreds of terabytes of vehicle data per day. The tools I created redefined the standard operating procedure for classifying and categorizing alerts. In many cases, the datasets I collaborated on and maintained became the gold standard and drove decisions across firmware, controls, and hardware teams.

My favorite tools and products involved visualization of datasets. Because the volume of telemetry was so large, creating a responsive and insightful dashboard often required multiple layers of aggregations before being consumed by custom visualization tools powering interactive dashboards. My self-service visualization suite became the de facto solution for anyone who could build or wanted to access and manipulate these large datasets.

My final contribution to Tesla was a complete self-service dataset builder. This project was a startup within a company. In the beginning, chance of success was very low and at Tesla being fired is always a risk. I shirked my other responsibilities and we slowly created a product too good to ignore. It leveraged custom streaming algorithms to identify and then compute metrics over time periods of interest and link temporal sequences of these events through a graph to classify and categorize complex events in the fleet. When I left it was being used by all firmware subsystem teams to automatically diagnose Robotaxi events. I suspect it will eventually be the standard method for classifying all service appointments, but I do not know which way it will evolve without me.

Anfa

While I truly enjoyed the challenges presented by Tesla, I eventually felt that I needed more variety in the problems that I solve and the skills that I develop. On the weekends, I started building tools and automation to help my friend’s venture capital firm source companies. This continued for a year while working at Tesla before I thought it might be interesting to take a more active role in evaluating companies and speaking to founders about their products. I decided to begin a full time role as a partner at Anfa. This chapter has just begun.