Work
YouTube
Programming Specialist, Youth & Learning · 2022 – Present
I report to the Global Head of Strategy & Programming and work at the intersection of human curation and machine learning systems.
My team was instrumental in transforming YouTube Kids from an unsafe, untrusted product into a platform parents can rely on. We did this by combining machine learning with systematic human evaluation — and the results were significant enough to be covered by the Wall Street Journal.
What I do:
- •Design quality evaluation frameworks for recommendation systems serving 10M+ weekly users
- •Curate preference data that shapes how models surface content — the "human feedback" in the loop
- •Own programming operations for priority markets (France, Germany), responsible for 1B+ views
- •Conduct safety evaluation for YouTutor, an unreleased conversational AI assistant
- •Lead the Common Sense Media partnership for YouTube Learning, piloting in 12 US school districts
Impact:
- •50%+ reduction in inappropriate content on YouTube Kids
- •45% increase in quality watch time
- •Built cross-cultural evaluation frameworks now used across EMEA
UC Berkeley Human Rights Center
Open Source Investigator · 2020 – 2021
Before tech, I worked on evidence verification for human rights documentation, collaborating with Amnesty International's Digital Verification Corps and The Washington Post.
This work was about establishing ground truth from messy, ambiguous data — collecting, verifying, and cataloging open-source evidence of human rights violations. I see a direct throughline to AI alignment: both require careful judgment about what's true, what's harmful, and how to make those distinctions at scale.
Education
University of California, Berkeley
BA in Global Studies · Minors in Political Economy and French
3.83 GPA, Cum Laude · 2017–2021
Sciences Po Paris
Exchange Program · Instruction in French · 2019–2020