Highlights:

  • The platform provides a detailed testing approach to uncover unexpected model behavior, including biases, false positives, and critical healthcare misclassifications.
  • The platform enables organizations to create precise test cases from their AI model’s datasets for evaluating effectiveness in essential scenarios.

Kolena Inc., a startup that tests machine learning and artificial intelligence models, recently announced that it has raised USD 15 million in new funding to support its mission to alter how AI systems are validated and tested fundamentally.

Kolena, an AI and ML testing platform, was established in 2021 and enables teams to test model behavior thoroughly. The platform empowers enterprises and teams to construct tailored test cases using their AI model’s underlying datasets, facilitating the assessment of model effectiveness in critical scenarios and applications.

Kolena contends that businesses have shown reluctance to fully adopt AI technology because there is no comprehensive solution for assessing and confirming AI systems’ reliability, equity, accuracy, and impartiality. Recent news stories highlighting prominent AI shortcomings have underscored the demand for a standardized and meticulous evaluation system for AI quality assurance, which would offer insights into a model’s effectiveness within the specific real-world situations in which it is designed to operate. This is the primary focus of Kolena’s efforts.

The company’s platform provides what is referred to as a “fine-grained” testing methodology that enables teams to identify unexpected model behavior, such as biased hiring algorithms, false positives or misclassifications in crucial healthcare applications, and other failure modes. The solution also helps teams and businesses launch their models more quickly by doing away with the need for time-consuming experimentation-based testing.

With numerous Fortune 500 companies, top AI startups, and governmental organizations among its clients, Kolena has experienced rapid growth.

Lobby Capital LLC led the Series A round, with participation from SignalFire Inc., Bloomberg Beta LP, 11.2 Capital LP, and Zero Prime Ventures LLC. The funding will be used for research and development to create the firm’s AI quality framework and toolings that enable businesses to create reliable AI systems.

Eric Carlborg, Founder of Lobby Capital, said in a statement, “It’s no secret that unpredictable behavior by AI and ML models is one of the key constraints holding back more widespread AI adoption. Enterprises cannot risk deploying a model as part of their tech stack — much less as part of a customer-facing application — without assurance that it is safe, reliable and fair.”

Carlborg continued that Kolena’s solution addresses fundamental problems by “bringing rigor and visibility to the testing process that goes far beyond ‘statistical accuracy’ while simultaneously transforming testing from an inefficient manual process into a true engineering discipline.”