W&B

Weights & Biases is a machine learning developer platform founded in 2017 in San Francisco by Lukas Biewald, Shawn Lewis, and Chris Van Pelt. The tools provide experiment tracking, model registry, evaluation, and production monitoring for deep learning and LLM teams.

W&B is used by more than 1,000 enterprise customers including OpenAI, NVIDIA, Meta, and Toyota, along with hundreds of thousands of individual researchers. Its Weave product focuses on LLM application observability.

CoreWeave announced the acquisition of Weights & Biases in March 2025 for approximately $1.7 billion. Prior to the acquisition, W&B had raised approximately $250 million across rounds led by Insight Partners, Coatue, Felicis, and Tiger Global.