Snorkel AI

Snorkel AI is an artificial intelligence company co-founded by CEO Alexander Ratner, Christopher Ré, Paroma Varma, Braden Hancock, and Henry Ehrenberg. Born out of research from the Stanford University AI Lab in 2015, Snorkel creates a data-centric AI platform, Snorkel Flow, that does not rely on manual labeling of data.

Snorkel AI relies on an approach the company calls "programmatic labeling," which allows users to programmatically label large quantities of training data by writing "labeling functions." This differs from conventional AI methods, which traditionally rely on hand-labeled machine learning techniques, rules-based approaches, black box models, or APIS.

As of 2021, Snorkel AI's Snorkel Flow platform has conducted case studies with several companies in a wide variety of industries, most prominently Google, Apple, Intel, Tide, as well as in the banking, government, healthcare, insurance, and telecom industries.

Snorkel AI has completed four funding rounds since 2019. In its most recent round, Snorkel closed an $85 million Series C round led by BlackRock and Addition Ventures on August 9, 2021 that valued the company at $1 billion.

Snorkel is a company offering a platform for building and manage machine learning training data that was founded by a team of researchers at Stanford in 2016. Snorkel offers their users a Python programming library to systematically build training sets, create weak supervision models, and train machine learning models.

Notable users and sponsors of Snorkel include: Google, Intel, IBM, DARPA, National Institutes of Health (NIH), Accenture, Alibaba.com, Linkedin, Stanford Medicine, NEC, Chegg, Facebook, SAP, Teradata, U.S. Department of Veterans Affairs, VMware, Toshiba, YouTube, NASA, FDA, Ant Financial, Office of Naval Research, Gordon and Betty Moore Foundation, American Family Insurance, United States of America Department of Energy, Hitachi, BASF, Infosys, and Microsoft.

The Snorkel Python library can be install for pip users by typing 'pip install snorkel' and for conda users by typing ' conda install snorkel -c conda-forge'. The library allows users to label training data with Labeling Functions (LFs), perform data augmentation with Transformation Functions (TFs), and monitor critical data subsets with Slicing Functions (SFs).