Full-access lesson
End of Volume Review & Interview Questions
A retailer has read the first eleven chapters and wants the τ-bench retail agent answering its customers by the end of the quarter: 10,000 conversations a day, peaks at three times the average, refunds and exchanges included. The team's default model is Claude Sonnet 5, for which nobody has published a τ-bench retail result. Its API account sits on the lowest standard tier. Six people are available to review refunds.
The complete curriculum
All 12 volumes, companion resources, interviews, and architecture reviews.
Executable engineering practice
Subscriber-only Python, Java, TypeScript, and SQL labs in the isolated runner.