The kinds of systems I build
Architecture matched to the job. Sometimes that means a lightweight internal tool. Sometimes it means a multi-agent architecture with evals, routing, and event-driven logic.
Internal tools
Custom dashboards, analytics, admin systems, and management interfaces built for specific workflows.
Client delivery systems
Systems that automate intake, routing, follow-up, and delivery. Turn manual handoffs into clean process.
Agent workflows
Multi-step AI systems that plan, act, log, and stay in bounds. Useful work, not theater.
Developer portals
API documentation, SDK onboarding, system status, and developer-first information architecture.
AI product prototypes
Fast prototypes that can grow into real products. AI features, automations, and conversational interfaces.
Production-minded MVPs
Systems that start small but are built to scale. Logging, observability, and clean architecture from day one.
Architecture Principles
Front end, back end, and workflow layers
Clean separation of concerns. UI, business logic, and orchestration each have their place.
APIs, orchestration, memory, and tool routing
Systems communicate via APIs. Agents have memory, tools, and routing logic that keeps them on track.
Debuggable systems with room to grow
Logging, tracing, and observability baked in. Easy to understand what's happening, easy to improve.