Case Study: Woven by Toyota accelerates continuous learning with Weights & Biases

How Woven Leverages Weights & Biases to Drive Continuous Learning

Woven by Toyota, the mobility technology subsidiary of Toyota Motor Corporation, needed a faster way to support continuous learning for autonomous driving. Its traditional “Autonomy 1.0” ML workflow for data collection, curation, labeling, retraining, and deployment was too slow, with the full process taking over a year and manual steps that did not scale. Woven by Toyota used Weights & Biases to help bring structure and speed to its machine learning development.

With Weights & Biases as a central system of record, Woven by Toyota improved experiment tracking, collaboration, and traceability across teams, while also using Sweeps, Reports, and Tables to automate workflows and create a de facto models leaderboard. The result was major efficiency gains, including 10x velocity in experiment tracking and a reduction in dataset curation time from 20 months to just 2 months. Weights & Biases helped Woven by Toyota share results faster and align teams around better ML performance.

Woven by Toyota

Evan Cushing

Machine Learning Engineer