How I Eliminated 19,494 Duplicate MongoDB Records (And What I'd Do Differently Today)

Short summary

A developer inherited a job platform where location-scoped listings spawned one MongoDB record per city — a USA-wide job created 19,495 duplicates and caused 7,000ms API response times. The fix involved collapsing records into a minimal representation while preserving location fidelity, without downtime or disruption to active users. The article walks through the diagnostic process and architectural remediation under real production constraints.

• •Job platform generated one MongoDB record per city for location-scoped listings, creating 19,494 duplicates per USA-wide job
• •API response times exceeded 7,000ms due to massive record bloat
• •Fix required schema redesign with zero downtime and provable performance gains

Generated with AI, which can make mistakes.