Weights & Biases raises $45 million in Series B funding

The startup says it now has more than 70,000 users at more than 200 enterprises

Weights & Biases, a startup building tools for machine learning practitioners, is announcing that it has raised $45 million in Series B funding.

The company was founded by Lukas Biewald, Chris Van Pelt and Shawn Lewis — Biewald and Van Pelt previously founded CrowdFlower/Figure Eight (acquired by Appen). Weights & Biases says it now has more than 70,000 users at more than 200 enterprises.

Biewald argued that while machine learning practitioners are often compared to software developers, “they’re more like scientists in some ways than engineers.” It’s a process that involves numerous experiments, and Weights & Biases’ core product allows practitioners to track those experiments, while the company also offers tools around data set versioning, model evaluation and pipeline management.

If you have a model that’s controlling a self-driving car and the car crashes, you really want to know what happened, Biewald said. “If you built that model years ago and you’ve run all these experiments since then, it can be hard to systematically trace through what happened” unless you’re using experiment tracking.

He described the startup as “an early leader” in this market, and as competing tools emerge, he said it’s also differentiated because it is “completely focused on the ML practitioner” rather than top-down enterprise sales. Similarly, he said that as machine learning has been adopted more widely, Weights & Biases is occasionally confronted by a “high-class problem.”

We’re not interested in selling to companies that are doing machine learning for machine learning’s sake, Biewald said. With some companies, there’s a mandate from the CEO to sprinkle some machine learning in the company. That’s just really depressing to me, to not have any impact. But I would actually say the vast majority of companies that we talk to really do something useful.