Skip to main content

Snorkel AI

Category: Data Infrastructure / Vector DBs

Snorkel AI provides a data-centric AI platform that enables enterprises to accelerate the development of AI applications by programmatically developing and managing training data. Snorkel AI was founded in 2019. The company is led by Alexander Ratner. Based in Redwood City, California, USA. Team size: 201-500. Total funding raised: ~$256M total (all rounds). Latest round: Series D ($85M, Sep 2024 led by SignalFire && Greylock). Key investors include ["Addition","Greylock","Lightspeed Venture Partners","GV (Google Ventures)","BlackRock","Prosperity 7 Ventures","QBE Ventures"].

Founded
2019
Headquarters
Redwood City, California, USA
Team size
201-500
Total funding
~$256M total (all rounds)

Value proposition

Dramatically reduces the time, cost, and manual effort required to build and maintain high-quality training datasets. Empowers organizations to build more accurate, adaptable, and robust AI applications by focusing on the data.

Products and solutions

["Snorkel Flow: An end-to-end platform for programmatic data development, including labeling, model training, analysis, and iteration.","Snorkel Evaluate: A comprehensive solution for building, running, and managing custom evaluations for AI models and agents.","Snorkel Expert Data-as-a-Service: A service that delivers expert-curated, specialized datasets for evaluating and fine-tuning frontier AI systems."]

Unique value

Snorkel AI pioneered the concept of programmatic data labeling, which allows domain experts to write labeling functions (LFs) to create vast, high-quality training datasets. This method is significantly faster and more scalable than traditional manual labeling, shifting the focus from hand-labeling data points to developing high-level rules and heuristics.

Target customer

Large enterprises and government agencies, particularly in regulated or complex domains like finance, healthcare, insurance, and the public sector, that require high accuracy and data privacy.

Industries served

["Financial Services","Healthcare & Life Sciences","Insurance","Government & Public Sector","Technology","Retail"]

Technology advantage

The core advantage is its weak supervision technology, which programmatically combines multiple noisy labeling signals to generate high-quality training data. This data-centric approach allows for rapid iteration on the data itself, which is often a more significant bottleneck than the model architecture, leading to faster development and more adaptable AI systems.

How they differentiate

Snorkel's primary differentiation lies in its programmatic, code-driven approach to data labeling (data-centric AI). This empowers in-house subject matter experts to create and manage training data, which is more secure, scalable, and adaptable than the manual annotation workforces or tools offered by competitors.

Main competitors

["Scale AI","Labelbox","Surge AI","Aquarium Learning"]

Key partnerships

["Google Cloud","Microsoft Azure","Amazon Web Services (AWS)","NVIDIA","Stanford University"]

Notable customers

["BNY Mellon","Chubb","Genentech","Wayfair","Memorial Sloan Kettering Cancer Center","U.S. Air Force and other government agencies"]

Major milestones

["Originated as a research project in the Stanford AI Lab.","Launched the open-source Snorkel project.","Secured a total of $237M in funding across multiple rounds.","Raised $100M in a Series D funding round at a $1.3B valuation.","Established partnerships with major cloud providers (Google Cloud, Microsoft Azure, AWS).","Launched Snorkel Evaluate and Snorkel Expert Data-as-a-Service."]

Growth metrics

Announced triple-digit revenue and customer growth. The latest funding round (Series D) increased the company's valuation to $1.3 billion.

Market positioning

Positioned as a leader in the data-centric AI movement, Snorkel AI provides an enterprise-grade platform for building and managing AI applications in-house. It is particularly strong in industries with complex data, high privacy requirements, and the need for deep domain expertise.

Geographic focus

Primarily North America and Europe, with a focus on large enterprise markets.

Patents and IP

The company holds multiple patents related to its core technologies, including programmatic labeling, weak supervision, and data-centric AI methodologies.

About Alexander Ratner

Alexander Ratner is the co-founder and CEO of Snorkel AI, and an affiliate assistant professor of computer science at the University of Washington. He completed his Ph.D. in computer science at Stanford, where he started and led the Snorkel project.

Official website: