Protege

Protege was co-founded by Bobby Samuels (CEO), Travis May (Chairman), Engy Ziedan (Chief Scientific Officer), and Richard Ho (CTO). The firm describes its leadership as combining experienced startup operators with depth across both data engineering and scientific research, though specific prior career details are not publicly detailed on the firm's materials.

Protege does not scrape data from the open web. It negotiates directly with data holders across industries — including hospitals, media archives, and industrial operators — and then curates, de-identifies, and structures that proprietary material into AI-ready datasets. The firm acts as an intermediary that aggregates supply from hundreds of sources and delivers it in formats tailored to each stage of model development.

Protege is not a raw-data broker. Its 'DataLab' team provides domain-specific curation, de-identification, and quality checks so that datasets arrive matched to a builder's specific use case — pre-training corpora, supervised fine-tuning examples, or uncontaminated evaluation benchmarks — rather than as generic data feeds.

The firm's public materials emphasize healthcare proprietary data, video archives, audio libraries, motion capture, and what it refers to as 'agentic data.' Healthcare appears to be the most developed vertical, with multiple announced evaluation benchmarks and partner collaborations in clinical documentation and medical billing.

Protege is a venture-backed operating company, not a family office. It has raised equity financing rounds including a $25 million Series A closed in August 2025 and a $30 million Series A extension led by Andreessen Horowitz announced in January 2026.