Anthropic Adds Andrew Tulloch to Its Inference and Performance Team
The AMW Read
A named researcher with experience across Meta, OpenAI, and Thinking Machines Lab joining Anthropic materially updates a leading lab's technical talent position.
Anthropic Adds Andrew Tulloch to Its Inference and Performance Team
Anthropic said Andrew Tulloch will join next week, according to reporting cited by AI Times. Tulloch is set to work on the lab's inference and performance team, focused on optimizing model training and user-response performance. He previously spent 11 years at Meta, joined OpenAI in 2023, contributed to GPT-4o, GPT-4.5, and o3 development, and later co-founded Thinking Machines Lab before returning to Meta.
The move is another high-profile reshuffling of frontier-model research talent, where individual technical leaders can affect both model capability and the efficiency of serving it at scale. For Anthropic, the emphasis on inference and performance is strategically consequential: improvements in training efficiency, latency, or response quality can translate into a stronger product experience and lower unit economics. It also follows recent reports of Anthropic expanding beyond its core Claude roadmap through an internal product incubator, making technical execution capacity increasingly important across the company.
Builders should treat the hiring signal as a reminder that model selection is not only about benchmark quality; reliability, speed, and cost behavior increasingly determine whether an AI feature can become a durable workflow. Investors should watch for evidence that talent migration is yielding measurable changes in Anthropic's model economics or enterprise adoption, rather than assuming that a prominent hire alone changes competitive position.
