Case Study: Mercor speeds AI talent matching with MongoDB Atlas and Vector Search

Mercor cuts profile load times from 13 seconds to milliseconds with MongoDB

Mercor, a rapidly growing AI-native talent network, faced significant infrastructure challenges as their platform scaled at 50% month-over-month. Their fragmented, relational database architecture resulted in 13-second candidate profile load times and a 6-hour lag before new candidates became searchable, which was unsustainable for their business. This hindered their core value proposition of fast, precise talent matching.

To solve this, Mercor adopted MongoDB Atlas and MongoDB Vector Search. This solution unified their search pipeline by combining semantic vector search with exact-match filtering in a single query on a consolidated document database. The results were transformative: profile latency dropped to milliseconds, internal search quality improved threefold, and the database scaled from 100 GB to 2 TB. This empowered Mercor's lean engineering team to support the company's growth to a $10 billion valuation while ensuring candidates become searchable instantly.

Mercor