AnthropicCo-Founder and Chief Architect
Jan 2021 — Present
Works on large-scale model training and AI research. Co-authored Anthropic research on helpful and harmless assistants using reinforcement learning from human feedback.
Sam McCandlish is a co-founder and Chief Architect at Anthropic, where he works on large-scale model training and AI research. Previously, he was a member of the technical staff at OpenAI and contributed to research on neural-network scaling and language-model training. He holds a PhD in physics from Stanford University.
Sam McCandlish is a co-founder and Chief Architect at Anthropic, where he works on large-scale model training and AI research. Previously, he was a member of the technical staff at OpenAI and contributed to research on neural-network scaling and language-model training. He holds a PhD in physics from Stanford University.
Career
OpenAIMember of Technical Staff
2018 — 2020
Contributed to research on gradient noise scale and neural-language-model scaling laws. Co-authored OpenAI research on scaling laws for neural language models.