Badge Post

My Meta badge on a laptop covered in stickers from my time at Meta.

If you’ve worked at Meta, you know how it works: when you leave, you write a badge post. It’s basically your goodbye letter. You list everything you shipped, thank everyone who carried you, and turn in your badge.

I just posted mine internally on Workplace, and wow, writing it made me realize how much can happen in just over a year. So here’s the external cut: what I did, what I learned, and where I’m headed.

By the numbers 📊

  • 1,192 diffs landed in production (~20 a week, every week)
  • 513,444 lines changed
  • 863 tasks closed
  • 290 Workplace posts (52K+ views, 921 reactions)

What I’m most proud of 🏆

  • Working on 0→1 internal tools. My first big project was an internal DevX troubleshooting tool, and sitting down with the engineers who used it and shipping fixes doubled usage, which taught me that talking to your users isn’t optional, even when your users are coworkers.

  • Leading our team’s AI push. I became the AI person: I ran all our AI trainings, built every one of our original skills from scratch, and got the team actually using Claude Code daily instead of just playing with it, which is a big reason I made IC4 in under a year, and at the AI hackathon I was named an outstanding performer and got a personal email from Zuck (still not over it).

  • Migrating our codebase to Rust. The most technically challenging and fun thing I did all year. Worth it.

  • Hosting student visits. I brought 140 UC Berkeley students through Meta SF to learn about engineering careers.

  • Becoming a Muse power user. I was an early power user of Muse, our new product, and one of the memes I made got retweeted by Alexandr Wang.

What this year taught me 💡

  • Culture is something everyone builds. This year had real ups and downs, including layoffs, and I saw firsthand how much a team’s energy shapes the work: the people who kept showing up with good vibes and kept helping each other made all the difference, and it’s a big part of why I picked Anthropic, where I want to help build a culture I believe in.

  • AI is a multiplier, not a replacement for craft. Leading our team’s AI push taught me that shipping fast is the easy part (I usually have 8 parallel agents at a time): the best results came when AI handled the boilerplate and humans still did thoughtful design reviews and mentorship, which left me more convinced than ever that engineering craft matters, and more curious about the questions behind the tools, which is a big part of why I’m joining Anthropic.

  • Being in person speeds everything up early on. As a new grad, nothing beats hallway questions, hearing senior engineers think out loud, and getting unstuck in five minutes instead of five hours.

  • Your first team matters a lot. At a big company, switching teams is harder than you’d expect, especially early on, so research your first team and push for the one you want, and don’t forget the side doors: hackathons, side projects, and cross-team relationships led to some of my favorite work this year.

Tips (aka what I wish everyone knew on day one) 📝

  • Track everything. The best advice I got was to put metrics on every surface you touch, because impact is the main thing that matters at a big company, and Meta taught me how to measure it and show it, including posting about my work: bragging on Workplace felt weird at first, but if you don’t tell people what you shipped, nobody will know.

  • Go somewhere with community. My new grad cohort was some of the smartest, kindest people I’ve ever met, they became my best friends, and you should pick a place where you’ll have people like that around you.

  • Find a mentor. I found one with 13 years at Meta, and it changed my whole experience: real code review comments, check-ins on how I was doing, caring whether I was learning or just closing tasks, and a steady push toward more scope.

  • Stay curious. Some of my best work came from things nobody assigned me, and the random experiments and side projects I picked up out of pure curiosity ended up saving my team real money, so don’t let a big company quietly make you feel like you should stay in your lane. Don’t. Wander.

What’s next 👀

After a year of leading AI adoption on my team, I kept coming back to the questions behind the tools: what these models should and shouldn’t do, who decides, and how we make sure they’re safe. At Anthropic, those aren’t side questions. Safety research, interpretability, and responsible scaling sit right next to product work instead of trailing behind it.

Which is why I’m joining Anthropic! I start October 13.

← Back to the blog