Agentic AI Developer designing autonomous LLM pipelines, multi-agent systems and RAG-powered knowledge bases that slot into real products — so early-stage startups can do more with less.
Eight years across design, full-stack engineering and now agentic AI — translating ambiguous product ideas into reliable, production-grade systems.
I build agentic systems for early-stage startups — autonomous LLM pipelines, multi-agent workflows and RAG-powered knowledge bases that ship into real products and deliver measurable results.
My path is unusual on purpose. I started as a designer, evolved into a UI engineer, then a senior software developer leading platform modernization, and today I design and ship agentic AI systems end-to-end. That arc gives me something rare: I think like a designer, ship like an engineer and reason about systems like an architect — all while staying close to the user and the business.
I work most days with LangChain, LangGraph and CrewAI for orchestration, OpenAI & Anthropic for reasoning, and Python + Next.js for everything around them. I’m happiest in the zero-to-one zone — turning a hard, fuzzy problem into a focused first version that works in production.
The fastest way to understand what AI can really do is to push past the demo and ship something. So I asked the model for a working FPS — once — and let the agentic workflow take it from prompt to playable build. It’s a small thing with a big signal: agents are ready to do real work, when you architect them right.
Play the gameA decade of compounding craft — each role chosen to widen the lens: brand, interface, full-stack, architecture, and now autonomous AI systems.
A consolidated stack honed across product engineering, design systems and applied AI — chosen for what actually ships and scales.
LLM agents that plan, retrieve and act across tools, APIs and long-running workflows.
Production retrieval pipelines, vector stores and prompt strategies for context-aware systems.
End-to-end product engineering across modern web and Python services.
AWS-native automation, schedulers and integration with enterprise data systems.
Design system thinking, conversion-led layouts and frictionless user journeys.
Real-time analytics, role-based admin tooling and operational dashboards.
Third-party platform integrations with reliability and idempotency in mind.
Comfortable owning ambiguous problems and shipping the first version into production.
“The best way to understand AI’s limits is to push them — then engineer past them.
— Operating principle
A computer science foundation, layered with focused, modern training in agentic AI and full-stack delivery.
If you’re building an AI agent, a RAG system, or a product that needs sharp engineering and clean design — let’s talk.