Four flagship case studies carry the portfolio. Everything else moves out of the way unless it earns its place with deeper proof.
Multi-agent orchestration with trust governance, approval gates, durable workflows, and the interfaces operators need when autonomy stops being a toy.
Production backend engineering under compliance pressure: long-lived ownership, careful rollout, and reliability work that had to hold up in real clinical operations.
A utility-grade developer tool that treats installation, runtime footprint, and operator clarity as product work, not afterthoughts.
Plugin-based orchestration surface with real-time dashboarding, offline-first task state, and a local bridge that makes agent tooling usable from software people already run.
Older experiments and weaker-fit projects stay reachable in a hidden archive so the main index can stay tightly aligned with the agent infrastructure story.