Burgers, TubiTV, and decision theory
It was a typical chilly San Francisco summer day when I met up with Marios for lunch at the now-closed Dirty Water restaurant inside the Twitter building. I don’t remember what we had, but knowing Marios, it was probably burgers.
It was one of the most fortuitous burgers I’ve had. At the time, my career was at a crossroads. Certainly, there were some highlights I could point to — I was the second major contributor to pandas; I co-founded (and sold) my own company; the GDP of small countries traded through the algorithms I wrote back at AQR. I was also comfortably leading a team and had a supportive manager, so I didn’t have to go anywhere. But internally I was craving for a change. I felt increasingly behind on the latest technology and I spent all of my fighting for resources to shield my team from getting randomized all the time. I wasn’t writing code or designing new features. The skills I was acquiring were not building towards a future that I wanted for myself.
Luckily, my conversation with Marios led to my joining TubiTV as V.P. of Engineering. It’s been a blast for the past year and change. Part of my job description is to grow the team. In the course of speaking with many potential Tubies, I’ve often drawn upon my own decision process to help them understand whether Tubi is the right choice for them. I’d like to condense them here before the passage of time makes me forget.
Tubi at a Glance
Last year, we announced that we surpassed 20 million users and have tens of thousands of titles on the platform. During this period, we’ve greatly expanded our distribution, our core services can handle 10x the scale by switching to Elixir/Scala, and our machine learning is far more sophisticated than the year before. But what makes me most proud is that all of it was accomplished with an extremely lean engineering organization that doesn’t come at the cost of work/life balance. As a point of reference, Hulu added 2.5 Tubi engineering teams in 2018 alone.
While we’ve come far in the past year, I’m even more excited about Tubi’s future. As the streaming wars heat up, fragmentation of originals will only increase subscription fatigue and drive viewers to free services. Millennials vastly prefer streaming to linear TV. And as the viewership shifts, so will the ad spend.
Tubi Engineering Culture
Of all the companies I’ve worked for, Tubi’s engineering culture resonates with me the most.
We value ownership above all. We look for engineers that are passionate about shipping quality software into production. We ask that they cultivate a deep sense of empathy for the user so that they’re the ones coming up with the next big feature, instead of just be handed specs, requirements, and deadlines. This means that we ask more from our team, but it also means that they have a lot more autonomy. It also allows us to cut down on process and do things like not having arbitrary deadlines.
We’re also an extremely passionate team. Marios and I both still spend substantial portions of our time writing code, designing services/features, or reviewing pull requests. We want to spend our time solving new and interesting technical challenges, so we prioritize people-time over machine-time.
We firmly believe that learning and growth are more valuable than tenure. Some of our senior engineering leaders started at Tubi straight out of school. Several of our senior engineers taught themselves how to code at Tubi. Everyone is highly encouraged to balance their own workload to make time for training courses, conferences, and other learning opportunities.
Why I joined Tubi
Chronologically, this section should have been first. But because this is the most personal, it’s also the most difficult to write.
Joining Tubi gave me exactly the combination of career growth I was looking for. I would be able to have a lot more impact on the company. I get to build a lot of new things from the ground up. I would get to write code again! And most importantly, I knew I’d fit in well with the culture here because everyone I met at Tubi took their work very seriously but their own egos much less so.
Of course, the decision wasn’t all puppies and flowers. Tubi isn’t a public company yet so there was no public valuation. It’s not dependent on VC money to survive, but paradoxically that meant it didn’t have the overinflated private valuations that most Silicon Valley engineers are used to. I did a lot of probabilistic analysis myself and certainly drove Marios/Farhad up the wall to play through scenarios, talk about the competition and outlook, etc. At some point I caught myself trying to build macroeconomic models about global markets, San Francisco's economy, and the growth of the tech sector. But all of my analysis was pointing to one thing: if all I cared about was maximizing expected monetary reward, I should go work for FAANG and stay there until retirement.
Fortunately for me, taking a temporary pay-cut wouldn’t really impact my lifestyle very much. I had the privilege of being in a position to adopt a regret-minimization strategy instead of a strategy that maximized only monetary reward. If Tubi wasn’t successful, I would have learned a metric f* ton, had a ridiculous amount of fun, and I could go back to BigCo in a year or two. If Tubi is successful in a few years, then certainly I would also reap the monetary rewards, as well as taking a big step forward in my own career.
May the Burgers Continue
Tubi hasn’t IPO’d yet, so I’m still mid-journey. But everything else I was looking for has been true, and more. I recognize the privilege of my position, but I’m not unique. Many engineers in Silicon Valley are now in the same boat as I was two years ago. I hope my own journey can be a helpful reference that helps them reach the best decision for themselves.