Organisation · company · US

Snorkel AI

Enterprise platform for "programmatic" data labelling — labelling training data with rules and weak supervision — plus expert-curated datasets and evaluators for frontier models.

Reference profile

Snorkel AI is an enterprise data company that grew out of the Snorkel research project at the Stanford AI Lab, which pioneered "programmatic" data labelling — using rules and weak supervision to label training data at scale rather than hand-annotating every example. Founded in 2019 by Alex Ratner and collaborators, it sells a platform for building datasets and evaluators for enterprise AI, and has more recently emphasised expert-curated data and benchmarks for frontier models and agents. It raised a $100m round in 2025 at a reported valuation of about $1.3bn.

Category
Data, training & services
Founded
2019
HQ
Redwood City, US
Funding
$100m Series D at a reported ~$1.3bn valuation (2025)
Key people
Alex Ratner