Deepchecks secures $14M for Machine Learning validation technology

Despite rapid growth, only half of machine learning models succeed in production, calling for enhancements to ensure success.

 Should we be wary of the risks that powerful Artificial Intelligence systems could pose to humanity? (illustrative) (photo credit: PEXELS)
Should we be wary of the risks that powerful Artificial Intelligence systems could pose to humanity? (illustrative)
(photo credit: PEXELS)

Given recent major advancements in machine learning, it is no surprise that the technology’s market has experienced notable growth in recent months, with a projected increase from $26 billion in 2023 to $226b. in 2030.

Despite this progress, only half of machine-learning models successfully reach production, and those that do, often encounter issues such as exceeding time and budget limits or experiencing significant failures. These challenges underscore the necessity for enhancements in the majority of machine-learning models.

Deepchecks, a graduate of Intel’s Intel Ignite accelerator program, aims to rectify that issue. The company has announced the availability of its solution for continuous machine-learning validation, having secured $14 million in seed funding.

Organizations are increasingly adopting machine learning and integrating it into their production processes, following software-like development cycles. While classical software development has established processes and tools for timely and quality project delivery over the course of several decades, machine learning presents additional complexities due to its multiple components and lack of transparency.

Advancing machine learning technology

To propel machine learning forward, testing and validation methodologies from software development can be applied. Deepchecks aims to assist practitioners, developers and other stakeholders by going beyond traditional Machine Learning Operations (MLOps) and providing visibility and confidence throughout the entire machine-learning life cycle, from development to deployment and operation in production.

“It makes little sense to deploy and monitor software that has not been thoroughly and systematically tested first,” Deepchecks co-founder and CEO Philip Tannor said in a press release. “Yet, this is what happens today with machine-learning applications. Deepchecks brings a new approach to MLOps, improving models by adding validation to every step of the machine-learning life cycle.”

Deepchecks co-founder and CTO Shir Chorev said: “Deepchecks introduces a community-driven MLOps framework that enables people, from data scientists and developers to C-level executives, to have a clear picture of how machine-learning applications behave from research to production.”

Yuval Rozio, the director of Alpha Wave Ventures, highlighted the importance of quality assurance in machine learning.

Quality assurance [QA] is often assigned to people who are in the beginning of their careers,” he said. “In machine learning, however, it’s often the most senior, highest-paid person in the room who is tasked with this. That’s because, as opposed to other domains, QA has not yet been systematized for machine learning, so it remains a dark art. Deepchecks is here to address this, moving machine learning and businesses forward.”