Wakeful AI was founded in Dubai in 2025 by fraud-risk and applied-ML operators who spent their careers explaining synthetic identity losses after the fact — and got tired of it.
In early 2025, our founders sat through a quarterly loss review where a single synthetic identity ring — 38 linked accounts, cultivated over 14 months across three institutions — accounted for more first-party loss than every other fraud vector in the portfolio combined. Every account had passed onboarding. Every account had a clean payment history right up until it didn't. Wakeful started as the model that would have caught it in month six.
CEO & Co-Founder. Formerly Head of Fraud Risk Engineering at a top-10 U.S. bank, where she rebuilt the synthetic identity detection program after a multi-ring bust-out. Started her career as a forward-deployed engineer at Palantir.
CTO & Co-Founder. Previously led graph ML infrastructure at a network-analytics firm serving payment networks. Built the entity-resolution systems that inspired Wakeful's cross-institution graph engine.
VP, Compliance & Model Risk. Fifteen years in bank model risk management, including leading SR 11-7 validation programs. Ensures every Wakeful model ships exam-ready on day one.
A model that helps you write a better loss report isn't a fraud tool — it's an accounting tool. We measure ourselves on losses that never happen.
A flag your compliance team can't defend to an examiner doesn't get acted on. Every model we ship is built to be validated, not just trusted.
Consortium intelligence should sharpen detection without centralizing customer PII anywhere outside the institution that owns it.