Inside the expert network training every frontier AI model | Garrett Lord

Handshake CEO Garrett Lord explains how the student career network turned its 500K PhD roster into a $50M ARR data-labeling business in four months.

• Handshake hit $50M ARR in the first four months of its data-labeling business and is on track to exceed $100M within 12 months.
• The core moat: zero customer acquisition cost versus competitors spending tens of millions monthly on LinkedIn recruiters and performance ads to find experts.
• Market shift is the key unlock: models got too good for generalist labelers, so labs now need domain PhDs — exactly who Handshake already has (500K PhDs, 3M master students).
• Handshake works with seven frontier labs simultaneously; Scale AI’s absorption into Meta created a trust/conflict vacuum that Handshake is filling.
• PhD contributors earn $150–$200/hour breaking frontier models versus $25/hour as teaching assistants.
• The new business was run as a fully separate org: separate team, office, engineering, finance, recruiting, all-hands, and compensation structure tied to hurdles.
• Lord argues AI creates Iron Man suits for junior employees rather than eliminating entry-level jobs, citing a first-day intern shipping a bug fix as evidence.

2025-08-24 · Watch on YouTube