Machine learning comppany Kolena announced that it has closed a $15 million Series B funding round led by venture capital firm Lobby Capital. Participating in the round are participation from SignalFire and Bloomberg Beta. This round brings Kolena’s total funding to $21 million. The funds will be used to expand the R&D team and the business.
Kolena is an ML testing platform that allows ML teams to rigorously test their model behavior. With Kolena, ML teams can identify and track failure modes to help engineering, sales, and customers better understand model performance and identify gaps. Kolena was founded in 2021 and is based in San Francisco, CA.
Kolena’s Impressive Testing Capabilities
Kolena was founded in 2021 by Mohamed Elgendy and others with a mission to build a model quality framework that provides unit testing and end-to-end testing for AI models. The Kolena platform identifies coverage gaps in model test data and also includes risk management capabilities to track potential risks of deploying AI systems.
Kolena allows users to customize test cases, evaluate model performance, and compare it to other models to find out why the model is underperforming. Rather than just looking at accuracy metrics, Kolena can test specific scenarios for AI product processing, providing more comprehensive model validation.
According to one survey, AI engineers spend only 20% of their time analyzing and developing models, and the rest of the time procuring and cleaning up data for training models. Another report found that due to the challenge of developing accurate, performance models, only about 54 percent of models ultimately move from pilot to production.
But there are other participants building tools to test, monitor, and validate models. In addition to established players like Amazon, Google, and Microsoft, a large number of startups are experimenting with new ways to measure the accuracy of models before and after they go into production.
Customer Privacy a Priority
According to the Kolena co-founder, the platform only stores model test results, and customers’ data and models are not uploaded, which helps to protect privacy. Kolena wants to be one of the few platforms that give customers full control over the elements of model testing.As enterprise AI applications continue to increase, the need to test and verify model performance becomes more and more prominent. Through its unique model quality framework, Kolena expects to provide AI teams with end-to-end model evaluation solutions. As a result, they reduce the risk of AI system failure or wrong output; and improving the overall quality of enterprise AI applications.
