Bio
PhD student at UCI IndyLab,
advised by Roy Fox.
Previously interned at Stripe (AI researcher, 2025), Duolingo (AI researcher, 2024), and Nvidia (2019).
HPI Fellow (Jan 2024 – Dec 2026).
Won the GPU Mode × Jane Street hackathon ($50k).
My current interest · 2026
World models for efficient test-time training: learning reward and feedback
signals in model-space, then using them to adapt RL policies and LLM agents
with less online interaction.
- Moonwalk: Inverse-Forward Differentiation · AISTATS 2026
- Learning to Design Analog Circuits to Meet Threshold Specifications · ICML 2023
- Reinforcement Learning Framework for Deep Brain Stimulation Study · IJCAI 2020
- Reinforcement Learning for Suppression of Collective Activity in Oscillatory Ensembles · Chaos 2020
Papers I found worth reading.
- Data Filtering Networks (Alex Fang, 2023)
The classic paradigm — better data → better models — is replaced by: better data → better filtering model → even better model.