Aaron Calvo
Founder & CEOForage AI, LLC
AIView Profile →Founder & CEO · Hebbia
George Sivulka is the founder and Chief Executive Officer of Hebbia, the New York AI company whose Matrix platform is used by investment banks, asset managers, private equity firms and law firms to run analysis across document sets too large for a person to read. He founded it in August 2020 while a PhD student at Stanford, left the doctorate to build it, and has raised more than $160 million for a company that was profitable before its largest round.
Last reviewed: Sep 19, 2026
Founder & CEO Hebbia
Last Reviewed: September 19, 2026
George Sivulka is the founder and Chief Executive Officer of Hebbia, the New York company behind Matrix — a platform built for the specific problem of doing serious analytical work across document sets no human team can read. A private equity firm running diligence, a bank preparing a deal, a law firm working through discovery: each faces thousands of PDFs, spreadsheets and decks, and each has historically solved it by paying junior staff to read them. Matrix decomposes a plain-language question into parallel sub-tasks, runs them across the whole corpus, and returns answers in a grid with citations back to source. Sivulka founded the company in August 2020 while enrolled in a Stanford PhD programme, left the doctorate to build it, and has since raised more than $160 million from Peter Thiel, Index Ventures, Andreessen Horowitz and Google Ventures — reaching a roughly $700 million valuation in July 2024 on a business that was already profitable.
Sivulka’s record is front-loaded to an unusual degree. He worked at NASA in his teens and published research on computational simulation that took a prize at an international conference before he had finished school. At Stanford he completed a Bachelor of Science in Mathematics with distinction in roughly two and a half years, then a Master of Science in Applied Physics, and entered a doctoral programme working on artificial intelligence.
He did not finish it. In August 2020 he founded Hebbia and left Stanford to run it. The founding observation was narrow and correct: search, as built, matches strings. Ask a document corpus about an “IPO” and it will not surface the paragraph that says “go public”. For casual search that is an annoyance; for a diligence team reading ten thousand pages under deadline, it is the whole problem. Hebbia’s first product attacked semantic retrieval directly, and Peter Thiel and Floodgate backed it in 2020.
Index Ventures led a $30 million Series A in 2022, by which point the company had begun the shift that defines it now — from retrieval to reasoning. The product became Matrix, which takes an analyst’s question, breaks it into parallel sub-tasks run across documents of unlimited length, and writes results back into a spreadsheet-like grid with citations. The interface choice matters commercially: financial analysts already live in grids, and Matrix meets them where they work rather than asking them to adopt a chat window.
The customer base concentrated in exactly the places where document volume is highest and error is most expensive. By mid-2024 Hebbia was serving roughly a third of the largest asset managers alongside investment banks, private equity firms and law firms, with BlackRock, KKR, Carlyle, Centerview Partners, Charlesbank, Ropes & Gray and Fenwick among the names. In July 2024 Andreessen Horowitz led a $130 million Series B, with Index Ventures, Google Ventures and Thiel participating, at roughly a $700 million valuation. Sivulka disclosed $13 million in annual recurring revenue and profitability at the time — fifteen times the revenue of eighteen months earlier, and a rare thing to be able to say while raising at that multiple.
In June 2025 Hebbia made its first acquisition, buying FlashDocs, a 2024-founded startup that automated generative slide-deck creation and was producing over ten thousand slides a day for enterprise customers. The logic was to own the last mile: a diligence answer is not the deliverable, the deck is. With FlashDocs the platform runs from raw document set to finished artifact without the analyst leaving it. Sivulka has since pushed Hebbia beyond finance into law and pharmaceuticals, where the underlying problem — too many documents, too little time, too high a cost of being wrong — is identical.
Stanford University
Bachelor of Science, Mathematics, completed with distinction in about two and a half years
Stanford University
Master of Science, Applied Physics
Stanford University
Doctoral study in artificial intelligence, not completed; left in 2020 to build Hebbia