Max Schwarzer, Vice President of Research heading Post-Training at OpenAI, has joined Anthropic as an individual contributor researcher focused on reinforcement learning. Schwarzer led the post-training work behind some of OpenAI's most important recent releases, including GPT-5 and the o-series reasoning models that introduced chain-of-thought reasoning capabilities.
Post-training — the process of refining a base model through techniques like RLHF, instruction tuning, and preference optimization — has become one of the most critical and competitive areas in AI development, and Schwarzer was at the center of OpenAI's efforts in this domain.
His decision to leave a VP-level role at OpenAI for an IC research position at Anthropic is a remarkable move that speaks to Anthropic's appeal to researchers who prioritize technical depth and alignment-oriented research. The transition follows a pattern of senior researchers with safety-adjacent expertise departing OpenAI for Anthropic. Schwarzer's reinforcement learning expertise is expected to directly contribute to improving Claude's reasoning and decision-making capabilities.
