In the final session of the SambaNova coding agent series, Kwasi Ankomah turns the improvement process into a scheduled production loop: trace, evaluate, diagnose, improve, and redeploy. He adds the safeguards production systems need — versioning, regression gates on the test suite, and human approval of changes — and runs the executor swarm at production scale on SambaNova.
WebinarUpcoming
Run self-improving coding agents in production: The full loop at scale
In the series finale, Kwasi Ankomah puts the whole improvement loop on a schedule and runs it at production scale on SambaNova — then judges the capstone coding agents built across the series.
Explore more events and resources
Browse our library of tutorials, webinars, and live sessions to keep building your AI skills.