10 Anthropic reinforcement learning researchers
Work on reinforcement learning for Claude ties together these Anthropic researchers and engineers, nine Members of Technical Staff and one Research Engineer, whose stated work ranges from the RLHF and Constitutional AI papers to RL environments, long context RL and RL scaling. Past roles at Apple, Twitter, Netflix and Character.ai give several of them a basis for comparison when an AI team is setting up post-training, RL environments or RL infrastructure.
Professionals to explore
Andy J.
LinkedInExperience: Member Of Technical Staff · Anthropic
$740 per 30-minute call
Language models, RL, scaling laws and safety are how Andy J. describes the work as a Member Of Technical Staff at Anthropic since October 2021. Andy was a Researcher at Andy Jones Consulting (2021), Founder of cartograph.ai (2019-2021) and a Researcher at GSA (2015-2019), a fit for questions on scaling laws and RL for language models.
Is this you? Claim your profile and get paid for calls.
Antoine Barge
LinkedInExperience: Member of Technical Staff · Anthropic
$740 per 30-minute call
Antoine Barge describes being a research engineer working on RL at Anthropic, as Member of Technical Staff since December 2025, right after Member of Technical Staff at Character.ai (2024-2025). Antoine was earlier a Staff Software Engineer at SliceX AI (2021-2024), Co-Founder at Various Startups (2019-2021) and a Research Assistant at Stanford University (2019-2021).
Is this you? Claim your profile and get paid for calls.
Evan Cox
LinkedInExperience: Member of Technical Staff · Anthropic
$740 per 30-minute call
Evan Cox supports teams focused on scaling reinforcement learning as a Member of Technical Staff at Anthropic since January 2024 and previously supported inference performance teams focused on efficiency and latency. Evan was Engineering Manager: Model Development, Inference, and GenAI Infrastructure at Netflix (2020-2024) and, at Airbnb, Machine Learning and Data Engineering Manager (2018-2020) and Software Engineer, Machine Learning (2017-2018).
Is this you? Claim your profile and get paid for calls.
Haochen Zhang
LinkedInExperience: Member of Technical Staff · Anthropic
$808 per 30-minute call
Haochen Zhang works on RL and post-training of Claude models as a Member of Technical Staff at Anthropic since April 2024. Haochen also holds a PhD in Computer Science from Stanford University (2019-2024) and was a Visiting Researcher at the Vector Institute (2018) and the Massachusetts Institute of Technology (2018), a fit for questions on post-training Claude models.
Is this you? Claim your profile and get paid for calls.
John Boccio
LinkedInExperience: Member of Technical Staff · Anthropic
$740 per 30-minute call
John Boccio works on RL for Claude as a Member of Technical Staff at Anthropic since March 2026, right after Member of Technical Staff at SpaceXAI (2024-2026). John was a Machine Learning Engineer (2023-2024) and Software/Machine Learning Engineer (2020-2023) at Apple and holds a Master of Science in Electrical Engineering from Stanford University, specialized in machine learning.
Is this you? Claim your profile and get paid for calls.
Mukund Sudarshan
LinkedInExperience: Member of Technical Staff · Anthropic
$740 per 30-minute call
Long context reinforcement learning is Mukund Sudarshan's area as a Member of Technical Staff at Anthropic since December 2025, right after Member of Technical Staff at Character.ai (2025). Mukund was Co-founder and CTO of Quilt Labs Inc (2023-2024), a Founder Fellow at South Park Commons (2022-2023) and a Visiting Researcher at Harvard Medical School (2016-2019).
Is this you? Claim your profile and get paid for calls.
Ruhua J.
LinkedInExperience: Member of Technical Staff · Anthropic
$740 per 30-minute call
Ruhua J. describes leading the Universes team in research at Anthropic and working on reinforcement learning with large language models to train Claude, as a Member of Technical Staff since April 2023. Ruhua was a Staff Machine Learning Engineer at Twitter (2017-2023), a Software Engineer at Akamai Technologies (2015-2017) and an HPC Engineer at the University of Connecticut (2014-2015).
Is this you? Claim your profile and get paid for calls.
Shauna Kravec
LinkedInExperience: Member Of Technical Staff · Anthropic
$875 per 30-minute call
Shauna Kravec has worked on reinforcement learning with large language models as a Member Of Technical Staff at Anthropic since January 2022, and has been Chief Executive Officer of Hofvarpnir Studios since 2021. Shauna was a Machine Learning Engineer at Clostra, Inc. (2021) and a Postdoctoral Researcher at the University of California, Riverside (2019-2020).
Is this you? Claim your profile and get paid for calls.
Xiaoyong Liang
LinkedInExperience: Research Engineer · Anthropic
$740 per 30-minute call
Xiaoyong Liang does RL env and post-training work as a Research Engineer at Anthropic since September 2025, and is also Cofounder and CEO of The Human Data Company (since 2025) and an angel investor in AI startups. Xiaoyong was a Tech Lead at Apple (2018-2023), working on spatial understanding for Apple Vision Pro, and Cofounder of Aesop Labs (2024-2026).
Is this you? Claim your profile and get paid for calls.
Yuntao Bai
LinkedInExperience: Member Of Technical Staff · Anthropic
$2,693 per 30-minute call
Yuntao Bai is lead author of Anthropic's RLHF and Constitutional AI papers and has been a Member Of Technical Staff at Anthropic since July 2021, right after Sherman Fairchild Postdoctoral Fellow at Caltech (2018-2021). That work makes Yuntao the pick for questions on RLHF and Constitutional AI methods.
Is this you? Claim your profile and get paid for calls.
Choose the right perspective
For RLHF and Constitutional AI questions, Yuntao Bai is the clearest pick, while Ruhua J., Haochen Zhang and John Boccio describe RL and post-training work on Claude itself, and Shauna Kravec reinforcement learning with large language models. Xiaoyong Liang covers RL env work, Mukund Sudarshan long context RL, and Evan Cox and Andy J. RL scaling; Antoine Barge and Mukund both offer a comparison with Character.ai.
Questions to take into the conversation
- 01How do you decide which tasks get their own RL environment, and how do you check that it rewards the behavior you want?
- 02What changes in RL training infrastructure as runs scale up, and where do performance bottlenecks show up first?
- 03When is human feedback worth its cost in post-training, and when do AI-generated feedback methods work as well?
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.