HyperAIHyperAI

Command Palette

Search for a command to run...

Former Meta Legal AI Researcher Seeks $6M for Legal Tech Startup

The legal AI boom is drawing talent out of leading technology labs, as Ethem Coban, a former researcher at Meta, has departed to found a legal technology startup. Coban, who spent nearly two years at Meta and most recently worked at Superintelligence Labs, helped train the company’s language models to reason through legal questions and reduce hallucinations. He left earlier this month and is currently in early-stage discussions to raise $6 million at a $25 million valuation, though the terms remain subject to change. Coban is expected to present his startup to investors as a provider of tools that allow law firms and corporate legal teams to build and customize AI models around their own proprietary knowledge. The concept is based on the growing belief that legal organizations may prefer to control their own models rather than rely on externally licensed systems, giving them greater authority over sensitive data and workflow design. His approach also taps into broader industry momentum, as high-profile legal technology companies such as Harvey and Legora have drawn strong demand and prompted wider investor interest in AI-driven legal automation. The startup arrives in an increasingly competitive market where major law firms are moving beyond pilots and committing serious resources to internal AI capabilities. Kirkland & Ellis, the world’s largest law firm by revenue, has pledged $500 million toward developing proprietary software, while Latham & Watkins has purchased its own Nvidia servers to customize models in-house. Against that backdrop, Coban’s competitive pitch will likely center on his technical experience inside a frontier model developer and his legal education at American University and Goethe University Frankfurt. His challenge will not be convincing legal businesses of AI’s value, but persuading them that he is uniquely equipped to help them build the technology themselves.

Related Links