Mark Chen (OpenAI)

Before joining OpenAI, Chen was a quantitative trader at Jane Street Capital, where he built machine-learning models for futures trading; he joined OpenAI in 2018.⁠[1]

On July 7, 2021, Chen, as first author with Jerry TworekJerry TworekJerry Tworek is an artificial intelligence researcher who worked at OpenAI for about seven years, served as its vice president of research, and took part in research on Codex and reasoning models such as o1. After leaving OpenAI in early 2026, he founded the AI lab Core Automation.Open the full entry, Heewoo Jun, Qiming Yuan and others, released the paper “Evaluating Large Language Models Trained on Code,” which introduced Codex, a GPT language model fine-tuned on publicly available code from GitHub, and noted that a distinct production version of Codex powers GitHub Copilot. In the paper, Codex solved 28.8% of HumanEval problems, compared with 0% for GPT-3 and 11.4% for GPT-J, rising to 70.2% with 100 samples per problem.⁠[2]

On September 25, 2025, a16zAndreessen HorowitzAndreessen Horowitz (a16z) is a venture capital firm founded in 2009 by Marc Andreessen and Ben Horowitz. It invests from the seed stage through the growth stage and provides operational support to its portfolio companies.Open the full entry released the podcast episode “From Vibe Coding to Vibe Researching: OpenAI’s Mark Chen and Jakub PachockiJakub PachockiJakub Pachocki is a computer scientist who has been chief scientist of OpenAI since May 2024.Open the full entry,” hosted by Andreessen Horowitz’s Anjney MidhaAnjney MidhaAnjney Midha is an entrepreneur, investor, and researcher, the founder of AMP, and a visiting scientist in applied physics at Stanford University. He co-founded Ubiquity6, led developer platform ecosystem work after the company was acquired by Discord, joined Andreessen Horowitz in 2023 to work on AI investing, and currently serves at the firm as a Venture Partner.Open the full entry and Sarah WangSarah Wang (a16z)Sarah Wang is a venture capitalist and general partner on the growth team at Andreessen Horowitz (a16z). She joined a16z in 2019 and leads growth-stage investments in AI, enterprise applications and infrastructure, having invested in companies including Character.AI, Wiz, Cursor, Temporal and Exa.Open the full entry, with Chen and OpenAI chief scientist Jakub Pachocki as guests.⁠[3][4]

The episode discussed how GPT-5 combines fast responses with long-horizon reasoning, how to measure progress once benchmarks saturate, why reinforcement learning keeps surprising skeptics, stability tradeoffs in agentic systems, how coding models change software development, and OpenAI’s longer-term bet on an “automated researcher” that generates new ideas with real economic impact; it also covered prioritizing compute, hiring focused “cave-dweller” talent and protecting fundamental research inside a product company.⁠[3]

At OpenAI Chen spearheaded the creation of the image-generation model DALL-E, worked on adding image recognition to GPT-4, and led development of Codex.⁠[1]

As chief research officer, Chen sets OpenAI’s research roadmap and long-term technical vision and shapes and manages its research teams. Describing how he works with chief scientist Jakub Pachocki, he said: “We're both researchers, we pull on technical threads.”⁠[1]

Chen has competed in international coding contests and coaches the USA Computing Olympiad team. Speaking of OpenAI’s models in programming competitions, he said: “We've watched them become better than me, better than Jakub.”⁠[1]

X: Public account: https://x.com/markchen90 (opens in a new window).⁠[3]