Noma builds tools to detect security issues with AI applications

Companies worry that their eagerness to use AI has left them vulnerable to cyber threats. According to a recent survey of more than 350 IT leaders, more than half of executives surveyed say the complexity of AI applications is weakening their organization’s cybersecurity posture. More than two-fifths of executives believe their security teams lack the skills needed to protect their AI applications and complex workloads, according to another survey.

The need for solutions to secure AI applications has spawned a whole host of startups. HiddenLayer and Protect AI, for example, focus on protecting AI systems from adversary attacks, while Cranium provides visibility into AI systems at the application level.

Today’s new startup, Noma Security, does all this and more, says founder and CEO Niv Braun. The startup develops tools to identify vulnerable data pipelines and code in areas of data science, as well as threats such as rapid injection attacks.

Braun launched Noma and Alon Tron after leaving client automation firm Verint, where he saw a “significant blind spot” in data security teams and the AI ​​lifecycle.

“The tools out there are designed for the software lifecycle,” Braun said. “However, data and AI lifecycles are inherently different. It involves unique R&D processes, relies on different technologies, and addresses unique vulnerabilities among AI models.”

Noma is designed to help detect – and fix – inaccuracies in components used to develop and run AI applications. For example, the platform scans “sensitive” data from model training datasets, such as text that contains personally identifiable information.

Braun says companies can be better equipped to deal with threats if they can consolidate AI application security functions into a single dashboard that can be deployed in cloud-based and self-hosted environments.

“Many single-function security solutions are flooding the market, trying to take advantage of the AI ​​hype,” he said. “By helping companies understand the foundation of their AI application and giving them control, Noma not only helps organizations reduce and mitigate risk, but also gives them the confidence they need to quickly expand their AI use and move into higher value areas of the business. .”

Noma, founded in 2023, already has paying customers, Braun said, including Fortune 500 customers in B2B software, financial services and retail. The startup recently closed a $25 million Series A funding round that Braun said was initiated by investors; Noma had some money left over.

“We decided to accept it [$25 million]even if there is still a lot of money in the bank, to speed up our development especially our teams that go to the market and our efforts,” he said, adding that Noma plans to double the number of the team of 20 people next year.

Ballistic Ventures led Noma’s Series A, bringing the company’s total funding to $32 million. Noma has raised $7 million in an undisclosed seed round led by Glilot Capital Partners in partnership with Cyber ​​Club London.