1 unread notifications

Noam Brown is a research scientist at OpenAI working on reasoning, reinforcement learning, self-play and multi-agent AI. He contributed to OpenAI’s reasoning models and previously helped develop CICERO at Meta and the poker systems Libratus and Pluribus with Carnegie Mellon collaborators.

Unclaimed profile

Noam Brown is a research scientist at OpenAI working on reasoning, reinforcement learning, self-play and multi-agent AI. He contributed to OpenAI’s reasoning models and previously helped develop CICERO at Meta and the poker systems Libratus and Pluribus with Carnegie Mellon collaborators.

Career

Research Scientist at OpenAI from 2023 to Present
Research Scientist at Meta / FAIR from 2018 to 2023
Research Assistant at Carnegie Mellon University from 2012 to 2018
Research Intern at DeepMind from 2017 to 2017
Research Assistant at Federal Reserve Board of Governors from 2010 to 2012
Algorithmic Trading Engineer at MJM Trading Group from 2006 to 2010

Investing Announcements

github.com

CICERO

Research code and models for Diplomacy agents combining language models and strategic reasoning, developed with the FAIR team.

Open ↗
Media appearance

Agent swarms, alignment and recursive self-improvement

Brown discusses multi-agent coordination, research acceleration and the challenge of evaluating alignment in a full interview with Dwarkesh Patel.

dwarkesh.com · Open ↗

Noam's Collections

Noam's Lists

In these lists

Cosigns Given