Gantry provides a continual learning and MLOps platform that connects to existing apps, data labeling services, and data storage to summarize and visualize data during training, evaluation, and deployment. The service helps ML teams decide when to retrain models and which data slices to use, and offers workflow tools to define metrics and compute them on specified slices. Founders Josh Tobin and Vicki Cheung (both formerly at OpenAI; Cheung was also a founding engineer at Duolingo) built Gantry to serve as a single source of truth for AI system performance and to surface issues such as bias even in unstructured data. The company positions itself alongside observability and monitoring tools but emphasizes continual learning and operational workflows beyond basic monitoring and explainability. Gantry emerged from stealth with $28.3M in combined funding (a $23.9M Series A and a previously undisclosed $4.4M seed) and says it will use the capital for customer acquisition and to expand its 22-person team. The company declined to disclose the number of users or customers.

Total raised

$28M

Funding rounds

2

Latest round

Seed

Latest activity

Jun 2022