Learn AI evals,
one card at a time
Five decks. 50 flash cards. Everything from what an eval actually is to running one in production — flip, swipe, repeat.
Foundations
What evals are — and what they are not
open deck →Types of evals
Every way to grade an output
open deck →Metrics & scoring
The numbers worth trusting
open deck →Build your first eval
From twenty prompts to a real suite
open deck →Tools & shipping
Evals that run without you
open deck →