The context
Renew Home wanted its engineering teams to use AI agents inside large, active codebases. The training took place across offices in Portland, New York, and San Francisco.
What Martian taught
Martian spent three weeks preparing a curriculum tailored to Renew Home's codebases. Morning sessions covered a repeatable approach to agent-assisted engineering, including context preparation, planning, verification, and model selection. We discussed the compute spent per correct result, including the retries and rework caused by a poorly matched model.
Hands-on work
In the afternoons, participants brought their own live tasks. Martian engineers worked with them as they applied the methods inside Renew Home's repositories.
