Raj Bakhru

About Raj Bakhru

Raj Bakhru is the global chief technology officer of Datasite, the virtual data room platform used by banks and private equity firms for M&A. By his count, about 60% of transactions over $100 million run through Datasite. He co-founded Blueflame AI in 2023, a vertical AI company for M&A that Datasite acquired in July 2025.

Bakhru began his career building research and trading systems at Goldman Sachs, Highbridge and Kepos Capital. He then founded a security firm, sold it to a private equity backed company and ran M&A and corporate strategy there before starting Blueflame.

In an AI Engineer conversation with Reducto co-founder Adit Abraham, he explains why Datasite buys document parsing instead of building it.

Career Highlights

• Global CTO at Datasite
• Co-founder of Blueflame AI (2023), acquired by Datasite in July 2025
• Founded and sold a security firm, then ran M&A and corporate strategy for the acquirer
• Built research and trading systems at Goldman Sachs, Highbridge and Kepos Capital

Notable Positions

On Build Versus Buy

With about 250 engineers and a target of $200 million in new revenue a year, Bakhru puts engineers on five revenue initiatives and buys layers that vendors solve every day, such as parsing with Reducto.

On Extraction Quality

Datasite's VDR bench asks a few hundred known-answer questions against a typical data room on almost every change, and must beat the same files in Claude or OpenAI projects. Bakhru credits better parsing with a 22% recall gain. See RAG.

On Proprietary Data

Blueflame joined Datasite for data nobody else had: Datasite's data rooms and Grata's database of about 20 million private companies.

Key Quotes

• "It's all edge cases. Like, literally, everything is an edge case." (on document parsing)
• "Most of the work actually is on the extraction side. You get that right and everything else can follow." (on RAG quality)

Related Reading

• RAG - Retrieval over parsed documents
• Model Routing - Balancing cost, speed and quality