Evan Reiser

Founder and Chief Executive Officer Abnormal AI

Last Reviewed: September 19, 2026

Executive Summary

Evan Reiser is the founder and chief executive of Abnormal AI, the security company he started with Sanjay Jeyakumar in 2018 on an argument that sounds obvious now and did not then: that the machine learning built to model human behaviour for advertising could model it just as well for defence. He had spent years doing exactly that at Twitter and at two adtech companies he co-founded, and he brought the technique across to a problem — email attacks that exploit how people behave rather than what software does — that the security industry had been treating as a filtering problem. Abnormal now says it protects millions of people and more than a quarter of the Fortune 500, and it raised $250 million at a $5.1 billion valuation in August 2024.

Career Highlights

• Founder and CEO of Abnormal AI since 2018, co-founded with Sanjay Jeyakumar.
• Raised a $250 million Series D at a $5.1 billion valuation in August 2024, led by Wellington Management with Greylock Partners, Menlo Ventures, Insight Partners and the CrowdStrike Falcon Fund participating.
• Had crossed $200 million in annual recurring revenue at that round, with more than 100% year-over-year growth.
• Led the company through its April 2025 rebrand from Abnormal Security back to Abnormal AI, the name it used at inception.
• Previously led product management and machine learning teams for the advertising business at Twitter.
• Co-founded Bloomspot, acquired by JPMorgan Chase, and AdStack, acquired by TellApart.
• Began his career at Eastman Kodak in research and development, working on machine learning for intelligence applications.
• Abnormal AI named to the 2026 CNBC Disruptor 50.

Professional Journey

Reiser took a BS in Computer Systems Engineering and Computer Science at Rensselaer Polytechnic Institute and began his career at Eastman Kodak, doing research and development on machine learning for intelligence applications — an unusual first job that put him on applied ML well before the field’s commercial boom.

The commercial phase came in advertising. From around 2010 he and his eventual co-founders worked on systems that analysed very large datasets, learned human behaviour, and made anomaly-based decisions in milliseconds, which is the operational description of programmatic advertising and, as it turns out, of behavioural security. He co-founded Bloomspot, later acquired by JPMorgan Chase, and AdStack, acquired by TellApart. He then led product management and machine learning teams for Twitter’s advertising business.

Abnormal’s own account of its origin is candid about what that background did and did not give them: they came to cybersecurity as outsiders, and saw what the industry’s insiders had missed — that the AI capable of modelling behaviour is also capable of stopping the attacks that exploit it. Business email compromise, vendor fraud, account takeover and the rest are not malware problems. They are impersonations of normal behaviour, and detecting them requires a model of what normal looks like for a particular person, a particular vendor relationship, a particular company. That is the thing adtech had spent a decade learning to build.

He and Jeyakumar founded the company in 2018. The name was Abnormal AI at inception, became Abnormal Security because the market at that point was not ready to treat artificial intelligence as the foundation of enterprise security, and returned to Abnormal AI on 16 April 2025 when the company judged that it was — a change that took the legal entity from Abnormal Security Corporation to Abnormal AI, Inc.

Between those two names the business scaled sharply. By the August 2024 Series D it had crossed $200 million in annual recurring revenue and was more than doubling year on year; the round raised $250 million at a $5.1 billion valuation, led by Wellington Management. The product surface has widened well beyond inbound email — outbound data-loss prevention, adaptive phishing simulation, identity, insider threat, and most recently detection of risky or malicious AI-agent behaviour in cloud environments, built with OpenAI models under a partnership announced in August 2026.

Reiser has also become one of the more prominent public voices on what AI does to the attacker’s side of the ledger, in Fortune and elsewhere — the argument being that the same tools lowering the cost of defence are lowering the cost of a convincing impersonation much faster.

Education

Rensselaer Polytechnic Institute

BS, Computer Systems Engineering and Computer Science