July 17, 2024


Unlimited Technology

Amazon vets land $10M for WhyLabs, a Seattle startup that monitors machine learning models

WhyLabs CEO Alessya Visnjic. (WhyLabs Photo)

The news: WhyLabs, a spinout from the Seattle’s Allen Institute for Artificial Intelligence (AI2), raised $10 million and released a new tool to support machine learning applications.

The problem: As more companies leverage machine learning and artificial intelligence, the need to capture and correct failures is becoming more urgent. Reliance on algorithms can lead to negative implications, as evidenced this week by Zillow Group, for example.

“The challenges begin once the machine learning system is live — it automates millions of decisions a day,” a WhyLabs spokesperson told GeekWire in an email. “Monitoring how well it’s working becomes critical, because machine learning systems fail in often catastrophic ways.”

The solution: WhyLabs’ software aims to prevent and diagnose in real time what Visnjic has described as “model meltdowns.” It offers an open-source library that integrates with various on-premise or cloud infrastructure services and monitors activity. The platform then uses complex statistical methods to spot any anomalies and alert engineers. WhyLabs is similar to Datadog, the publicly-traded cloud software monitoring company, but with a focus on back-end machine learning and AI processes.

The new tech: WhyLabs launched a new tool in October, AI Observatory, for companies to monitor, understand and improve their AI applications. Users can self-onboard the tool as a layer to a machine learning application. “We are removing barriers for the adoption of this essential technology,” said CEO and co-founder Alessya Visnjic in a statement. More than a dozen organizations used the tool in its first week.

The competition: When WhyLabs spun out of AI2 last year it joined a growing number of startups tackling different parts of the machine learning development and deployment process. TruEra and Fiddler AI, for instance, have similar functionality, but they target uses in specific industries.

The people: Visnjic previously spent eight years at Amazon helping the tech giant develop its machine learning infrastructure, and won Startup CEO of the Year honors at the GeekWire Awards this year. Two other co-founders, Andy Dang and Sam Gracie, are also Amazon veterans, and co-founder and COO Maria Karaivanova was formerly a Cloudflare executive and principal at Madrona Venture Group.

The backers: The Series A round was co-led by by Defy Partners and the AI Fund, helmed by Coursera co-founder and chair Andrew Ng. Existing backers, including Madrona and Bezos Expeditions, also participated. The $10 million Series A round builds on a $4 million raise last fall.

The future: WhyLabs plans to add features such as increased support for image, audio and natural language processing uses and increased integration with other tools. The 18-person company, with satellite offices in Raleigh, N.C., and Novi Sad, Serbia, aims to double its headcount in 2022. “We are a remote-first company. However, we value face-to-face time with the team,” said the spokesperson.

Source News