Plan, build, and improve background-agent infrastructure that enables reliable, trackable, and repeatable automated engineering work.
Develop systems for evaluating agent performance using task outcomes, quality measures, human intervention, cost, and other signals.
Create evaluation sets and feedback loops that help AI systems improve over time through regular review, tuning, and correction.
Improve how engineers discover, understand, and use shared AI tools and workflows.
Evaluate AI development tools, including code generation, AI-assisted review, and testing, against real engineering needs, and help integrate the ones that prove useful.
Develop visibility into which tools and workflows are useful, where they create friction, and how they should evolve.
