From OpenAI to Google: Here’s Why

10 years later, I’m back at Google to make real-world AI better

I recently made the decision to leave OpenAI and rejoin Google to work on pushing the frontier for enterprise.

Why now?

Foundational models are only getting smarter and smarter. Yet, it’s easy to spot real world disconnects. Models are winning gold medals at the International Mathematical Olympiad (an awesome milestone!), but still can’t complete my taxes for me.

I’m particularly interested in the research question of how to tackle real-world use cases that remain out of reach. Some major aspects that I’m keeping an eye on are agents, evaluations, and fine-tuning.

Summer 2025 marks one decade since the first time I joined (a very different) Google. Both me and the company have grown and changed a lot in that time.

Why Google?

Google is one of those tech giants, which is now impossible to describe in monolithic terms. Different corners and teams will have different vibes. After my first tenure at Google, I made friends and connections who now span many orgs, and their perspectives will almost certainly vary from the one I’m about to share.

For me, I’m returning to work with a team consisting of people whom I actually met at OpenAI. Tech is a small world.

I’m really grateful to have leaders who believe in me and who championed my return. I owe a lot to these folks. This unique role will also give me research latitude to work on meaningful problems that will improve Gemini models.

People matter a lot, and have always been a major driver of how I’ve decided to join teams or companies.

I am joining a fast-moving team full of brilliant people to rapidly make progress on this front. So far, I’ve already witnessed new milestones and cool demos. I can’t wait to share more soon.

My niche will be advancing the Gemini models to improve them based on real-world use cases. I have been awed with how much better models have gotten and what records were set on recent benchmarks. I also think these benchmarks are incomplete and there are many interesting research questions to figure out. We still need to determine what metrics we care about and how we hill-climb on them.

Google has a powerful distribution as well. I enjoy seeing a clear line from my work to impact. Google has one of the best networks to bring technology to enterprises, so I’m hoping to leverage this to translate AI wins back for good.

Industry Movement

The industry has seen a lot of talent moves and team reshuffling lately. There are a variety of reasons for this, but I also think this is a natural consequence as AI grows more powerful. And the pace of progress is accelerating (which is also something we need to take seriously for reasons well beyond role changes!).

Impact that used to take years can be had in weeks. As a result, I think we’ll continue to see folks move teams or companies more frequently than before.

I think competition is a good thing. As a personal consumer, I still use ChatGPT, and I get a kick out of using certain features and thinking of my friends who built them. During my time off, I used Cursor with Claude. There’s a lot of cool work and people out there, and I’m glad there are many industry solutions.

With this latest set of staffing changes, I now have friends and people I really respect at all the major labs and startups. I’m keen to see what they do next!

Giving myself a summer break

Not only has it been ten years since I rejoined Google, but its also about the same duration since I took a break.

The last time I had an extended period of time off was the summer of 2015. After graduating college, I traveled the world with friends and family, and prepared to move to SF. It was a super fun and memorable summer.

I highly recommend taking a month off between jobs. For me, that was the perfect balance of enjoying the break without getting too bored. Intentionally planning that in was one of the best decisions I made as part of the move.

This time I had about a month off between leaving OpenAI and starting at Google. I swam, I ran, and I meditated. I popped over to have tea with my family whom I am fortunate to live nearby in DC. I went on long walks in the summer evenings, relishing the long days. I enjoyed my morning coffee without any rush to open a laptop or get to the office.

I picked up new hobbies, too. I started angel investing, so I can stay connected to the startup world.

I know I’m very fortunate that I was able to intentionally plan a month off, and if that’s a luxury you can take, I’d encourage it.

Now I’m excited to jump in and get back into the thick of things.

Disclaimer: as always, all views are my own.