I’m a researcher currently at OpenAI, working on reinforcement learning and synthetic data; some models I’ve contributed to are 4o-mini, o*-mini, and o3.

I graduated from UC Berkeley, where I was advised by Pieter Abbeel and Igor Mordatch.

Email: matches my arxiv papers

Experience

2024-Present — OpenAI

I study reinforcement learning and synthetic data.

Some of my past contributions include:

• o3 & o4-mini
• 4.1-nano & 4.1-mini
• o3-mini
• o1-mini
• 4o-mini, audio

2022-2023 — HRT AI Labs

I worked on research studying deep learning for high-frequency trading.

2021-2022 — Facebook AI Research

I worked on research studying how we could scale Decision Transformer with new capabilities, such as multi-task, in-context learning, and online exploration.

2018-2021 — UC Berkeley

My research focused on the universal power of sequence modeling and how to leverage offline data to for reinforcement learning.

• First author of Decision Transformer and Pretrained Transformers as Universal Computation Engines
• Supporting author on several other papers (Google Scholar)
• Other highlights: Towards a Universal Decision Making Paradigm

I also spent a significant portion of my undergrad life teaching.

• Head TA for EECS 126 (Probability and Random Processes) in Fall 2020 and Spring 2021
• TA for EECS 126 in Fall 2019 and Spring 2020