10 OpenAI reinforcement learning researchers
Reinforcement learning appears in the OpenAI titles or work descriptions of all of these researchers, from RLHF for ChatGPT and LLM post-training to exploration problems, reasoning and the foundations of RL. Two have been research scientists there since 2020, two hold lead titles, and six are members of technical staff whose other titles include Microsoft, Meta, Google DeepMind and Genentech roles. They help teams planning RL or post-training work for language models.
Professionals to explore
Adrien Ecoffet
LinkedInExperience: Research Scientist · OpenAI
$2,693 per 30-minute call
Adrien Ecoffet has been a Research Scientist at OpenAI since 2020 and describes the work as reinforcement learning research, with a paper on Reinforcement Learning Under Moral Uncertainty. Adrien was Research Scientist (2019–2020) and AI Resident (2018–2019) at Uber, and Staff Software Engineer at Quora (2017–2018). Adrien fits questions on RL with uncertain or conflicting objectives.
Is this you? Claim your profile and get paid for calls.
Daniel A. Roberts
LinkedInExperience: Lead (Frontiers) · OpenAI
$2,693 per 30-minute call
Lead (Frontiers) at OpenAI since 2026, Daniel A. Roberts took that title after Lead (Foundations of Reinforcement Learning) (2025–2026), Technical Lead (Science of Reinforcement Learning) (2025) and Member of Technical Staff (2024–2025). Daniel is also a Visiting Scientist at the Center for Theoretical Physics at MIT (since 2020). Daniel fits questions on the science and foundations of RL and on how a research direction is set.
Is this you? Claim your profile and get paid for calls.
Ilya Kostrikov
LinkedInExperience: Member of Technical Staff (Research Scientist) · OpenAI
$1,885 per 30-minute call
At OpenAI since 2023 as a Member of Technical Staff (Research Scientist), Ilya Kostrikov works on RLHF research and deployment for ChatGPT, including LLM post-training, evaluations, tool use, frontier model performance prediction and user retention optimization. Ilya was Postdoctoral Scholar at University of California, Berkeley (2021–2023) and Research Associate at Google (2020–2021). Ilya is the pick for questions on taking RLHF from research into a deployed product.
Is this you? Claim your profile and get paid for calls.
James R Lee
LinkedInExperience: Member of Technical Staff · OpenAI
$1,683 per 30-minute call
Professor of Computer Science at University of Washington from 2006 to 2025, James R Lee is a Research Lead at OpenAI (since 2025) on core STEM and SWE reasoning capabilities and a Member of Technical Staff (since 2024) working on reinforcement learning, synthetic data and autonomous research. James was also Senior Principal Researcher at Microsoft (2024). James fits questions on RL for STEM and SWE reasoning and on synthetic data for training.
Is this you? Claim your profile and get paid for calls.
Joost Huizinga
LinkedInExperience: Research Scientist · OpenAI
$841 per 30-minute call
Joost Huizinga, a Research Scientist at OpenAI since 2020, has focused on sparse and deceptive problems in reinforcement learning, with progress on the exploration benchmarks Montezuma's Revenge and Pitfall. Joost was a Research Scientist at Uber (2018–2020) and a Ph. D. Student at University of Wyoming (2013–2018). Joost is a fit for questions on exploration and sparse rewards, where reward is rare or misleading.
Is this you? Claim your profile and get paid for calls.
Kevin Wang
LinkedInExperience: Member of Technical Staff · OpenAI
$1,683 per 30-minute call
Since 2025, Kevin Wang has been a Member of Technical Staff at OpenAI, doing research on RL and reasoning. Kevin was an ML Researcher at Princeton Language and Intelligence (2023–2024), a Quantitative Research Intern at Citadel Securities (2024) and an ML Engineer at Zoom (2023). Kevin suits questions on RL and reasoning research, and on moving from ML research at Princeton Language and Intelligence into OpenAI.
Is this you? Claim your profile and get paid for calls.
Masatoshi Uehara
LinkedInExperience: Member of Technical Staff · OpenAI
$1,683 per 30-minute call
For questions on applying reinforcement learning and reasoning models to biology and life science problems, Masatoshi Uehara is the one to ask: a Member of Technical Staff at OpenAI since 2026 working on RL and reasoning for life science, and a Research Scientist at EvolutionaryScale (since 2025). Masatoshi was a Researcher at Genentech (2023–2025).
Is this you? Claim your profile and get paid for calls.
Sivakanth Gopi
LinkedInExperience: Member of Technical Staff · OpenAI
$1,683 per 30-minute call
Sivakanth Gopi has been a Member of Technical Staff at OpenAI since 2025, working on reinforcement learning, and holds a PhD in Computer Science from Princeton University. Sivakanth was Principal Researcher (2023–2025), Senior Researcher (2020–2023) and Postdoctoral Researcher (2018–2020) at Microsoft. Sivakanth fits questions on RL research at OpenAI after several research titles at Microsoft.
Is this you? Claim your profile and get paid for calls.
Tai Vu
LinkedInExperience: Member of Technical Staff, Research · OpenAI
$1,683 per 30-minute call
LLM post-training, reinforcement learning and agent research are Tai Vu's specialties as Member of Technical Staff, Research at OpenAI since 2026. Tai was a Senior Machine Learning Engineer at Meta (2022–2026), building and scaling AI systems, and an Undergraduate Student Researcher at Stanford Artificial Intelligence Laboratory (SAIL) (2020–2021). Tai suits questions on RL for agents and on the engineering needed to run post-training.
Is this you? Claim your profile and get paid for calls.
Wei Hua Xiong
LinkedInExperience: Member of Technical Staff · OpenAI
$1,683 per 30-minute call
In 2026 Wei Hua Xiong became a Member of Technical Staff at OpenAI on RL Frontier, and is a PHD Candidate at University of Illinois Urbana-Champaign (since 2023), researching reinforcement learning for LLM post-training, training dynamics and the math behind these methods, including iterative rejection-sampling fine-tuning. Wei Hua was a Research Scientist Intern at FAIR @ Meta (2025) and Student Researcher at Google DeepMind (2024–2025). Wei Hua suits questions on post-training algorithms.
Is this you? Claim your profile and get paid for calls.
Choose the right perspective
For RLHF and post-training of language models, Ilya Kostrikov, Tai Vu and Wei Hua Xiong are the closest matches. Daniel A. Roberts and James R Lee hold lead titles and fit questions on research direction and reasoning, Adrien Ecoffet and Joost Huizinga fit exploration and classic RL problems, and Masatoshi Uehara is the one for RL applied to life science.
Questions to take into the conversation
- 01How do you decide whether a capability should come from reinforcement learning or from more supervised fine-tuning data?
- 02What makes a reward signal for a language model trustworthy enough to train against without the model gaming it?
- 03How do you evaluate whether an RL post-training run improved reasoning rather than just benchmark scores?
About this directory
This is a professional research starting point based on business profile data retrieved on . Titles and companies reflect that source snapshot; roles marked former had ended by then. Check the linked profiles for current details. Inclusion does not imply Instant Expert membership or availability.