Used-equipment marketplace · E-commerce · TAI Teams
MPB
Production AI evals for a marketplace team
Team size
20–40
product & ops
Programme length
2 days
live cohort
Focus
AI evals
shipped into production
Challenge
Where they started
AI-powered listing tools were shipping without a way to measure whether model quality was improving or regressing.
Approach
What we ran
- Two-day live cohort with the product and operations team
- Built a working evaluation framework for their AI listing tools
- Defined quality measures against real marketplace listings
- Wired continuous improvement loops into production