As companies rely more and more on machine learning models, they must constantly check the models to ensure they are working as expected and that the data is valid, unbiased and not corrupted in any way.
deep check, an Israeli startup has come up with a solution to continuously test models from development to production.Today, the company announced a $14 million seed investment and open source The version of the solution.The company also owns commercial productDeepchecks Hub, which began testing earlier this year.
CEO Philip Tannor and CTO Shir Chorev spent time in the Israeli army, where they worked on machine learning. Chorev said they saw first-hand the power and limitations of the technology, which is why they decided to start a company to work on QA models.
“So Philip and I are in a position to lead AI research and bring models into production. We both see the enormous value potential of algorithms in AI in general, but beyond that we also note that they do face some pretty significant challenges. challenges, especially being able to actually do what they’re supposed to do and do it right,” Jolev told TechCrunch.
She says it’s hard to know what’s going on inside these models without tools like Deepchecks, which can help data scientists make sure the data they use isn’t biased, misinformed, or based on corrupted input data—and that, over time, Usually well behaved.
As practitioners, what she and Tannor realized was missing was a tool that could help keep these models in check throughout the process to test and validate different aspects of the models, such as the integrity of the data. So, like any good entrepreneur, they built the tool they wanted and open sourced it. Recently, they also built a commercial version of the product, which provides standard enterprise features such as strong security and authentication.
“We bring a fair amount of understanding and practical experience in the field, and also know what the problems are and what we want,” she said. These experiences are relevant and helpful for them to build companies. The company reported that the open-source product has been downloaded more than 500,000 times and is used by companies such as AWS, Booking.com and Wix, as well as regulated companies they cannot publicly name.
Launched in 2020, the company took about a year and a half to roll out the first version of the open-source tool.
Chorev admits she’s a relative rarity as a female co-founder, but says she’s always been interested in the male-dominated space, and it’s been a driving force throughout her career. “Personally, I did grow up feeling like it wasn’t going to stop me. Instead, it was like okay, it’s harder, so I’m going to push and get it,” she says.
“I mean, it’s a challenging route no matter which way you choose, but I do try to help other people encourage people, whether they’re women or men, to take this route if they want to .”
Today, the company has about 15 employees and plans to increase that number by 25-30% this year. She cares about building a diverse workforce, though she admits it can be a challenge for a young company. “It’s a huge challenge and we’re really trying to encourage diversity right now and move forward as we grow,” she said.
Today’s funding round is actually a combination of two funds, a $4.4 million seed round at launch and a $9.5 million closing late last year. The investment was led by Alpha Wave Ventures, with participation from Hetz Ventures and Grove Ventures.