Intel / video

How We Built an Agent That Improves Itself — Zubin Aysola, Weights & Biases

Key takeaways · AI-distilled

In a live demo, W&B's ARIA turns a production trace into an offline task, finds the bug (a missing SDK call), writes a fix and benchmarks the new version against the one in production.
Aysola says benchmarks, evals and agent configs all change together, so measurement has to be airtight; his team keeps the production and research agents byte-for-byte identical.
Agent variants are defined in YAML so many can run side by side in an unconstrained , and each is scored two ways: pass/fail and relative comparison.
The suite runs 886 tasks, including simulated multi-turn users. Every production miss, and every production win, becomes a new task, so the team spends its time improving the system instead of writing benchmarks by hand.

Terms in this piece · Glossary

AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
sandbox — An isolated environment where AI-generated code or agent actions run without being able to touch anything real.