Snorkel AI: This Stanford AI Lab Spinout Secures $15 Million To Make AI Practical

By Amit Chowdhry ● Jul 17, 2020
  • Snorkel AI — a company that is working on making artificial intelligence practical — announced that it has launched out of stealth with $15 million in total funding

Snorkel AI — a company that is working on making artificial intelligence practical — announced that it has launched out of stealth with $15 million in total funding from Greylock, GV, In-Q-Tel, and several others. The company’s end-to-end Machine Learning (ML) platform Snorkel Flow enables developers and non-developers to build and deploy AI applications in a fraction of the time by programmatically labeling and managing the “training data” that fuels modern artificial intelligence. And paying customers using Snorkel Flow includes major U.S. banks, government agencies, and other large enterprises.

The process of manually managing, building, and labeling large training datasets can be a tedious process and it is considered one of the most significant bottlenecks to the adoption of AI. This process often requires weeks or months of manual effort for each application.

The Snorkel AI founding team – Alex Ratner, Chris Ré, Paroma Varma, Braden Hancock, and Henry Ehrenberg – saw how this training data issue was becoming the key problem in AI while working at the Stanford AI Lab.

After the team spent four years developing and deploying technology to solve this problem with Google, Intel and Apple, and organizations like DARPA and Stanford Hospital, they spun out of Snorkel AI and build an end-to-end platform that made this technology accessible to all enterprises.

Snorkel Flow is essentially a first-of-its-kind machine learning platform that uses a novel programmatic approach to building and labeling the “training datasets” that fuel modern artificial intelligence. And users can drive the end-to-end development process without spending months manually labeling and managing data.

Users develop “labeling functions,” (or rules or heuristics) and other programmatic operators — which Snorkel Flow automatically integrates to train state-of-the-art machine learning models. And users can easily improve and adapt these models just by editing their programmatic training data in Snorkel Flow’s guided interface.

Snorkel AI’s customers have already saved months of time and they are applying artificial intelligence to new problems that they couldn’t tackle before. For example, a major U.S. bank uses Snorkel Flow to quickly build artificial intelligence applications that classify and extract information from their loan portfolio, including for a recent time-sensitive use case that the bank had estimated would have taken months of manual labeling efforts. And with Snorkel Flow, the team produced a solution that was over 99% accurate in under 24 hours, and that could be quickly and easily adapted to new problems and business lines.

Key Quotes:

“Despite spending billions of dollars on AI, few organizations have been able to use it as widely and effectively as they want to. This is because available solutions either ignore the most important part of AI today – the labeled training data that fuels modern approaches – or rely on armies of human labelers to produce it. Our end-to-end platform, Snorkel Flow, focuses on a new programmatic approach to the training data that enables enterprises to use AI where they couldn’t before.”

— Alex Ratner, CEO of Snorkel AI

“Snorkel Flow is the first end-to-end ML platform that focuses on the data, making AI a reality for enterprises. We’ve consistently heard from Fortune 500 CIOs that they have been disappointed with their progress using AI, largely because they get stuck on the data. Customers’ rapid success with Snorkel Flow is a testament to the power of this new, data-centric approach, which has the potential to democratize AI across the enterprise. We are thrilled to partner with the Snorkel team as they drive this important market shift.”

— Saam Motamedi, Partner at Greylock and Snorkel Board Member

“The time, expertise, and costs involved in labeling training data present significant challenges to the U.S. government in applying AI to missions of national security. Snorkel AI provides a revolutionary capability that can greatly reduce the level of effort required to develop mission-ready machine learning models by addressing this critical data problem.”

— A.J. Bertone, a Partner at In-Q-Tel