Cloud (AWS)
Architecture, cost optimization, and platform engineering on AWS — the ground the AI is built on.
AI & Agent Systems · Software Architect
I help companies ship AI systems that survive production.
Fifteen years building production systems — cloud-native infrastructure, distributed backends, and APIs. Today I focus on the hard parts of moving AI from prototype to production: LLM applications, RAG pipelines, and multi-agent architectures built to hold up under real load.
Services
What I'm hired to do.
One specialty, three supporting capabilities. The cloud and platform work isn't a side offering — it's the production engineering that lets the AI actually ship.
Lead specialty
AI Engineering
Agentic AI, LLM applications, RAG, and multi-agent systems — taken from prototype to dependable production system.
Agentic AILLM appsRAGMulti-agentAI automationGenAI productsBedrock AgentCore
- From prototype to production
Reliability, evals, and guardrails so AI holds up with real users. - Built on real infrastructure
Kubernetes, AWS, and the MLOps discipline to operate it. - Independent and senior
You work directly with the architect — no handoff, no layers.
DevOps & MLOps
Kubernetes, CI/CD, observability, and ML deployment — the discipline that keeps production AI reliable.
Full-Stack & Distributed Systems
Backend systems, APIs, and distributed architectures — with frontend when an engagement calls for it.
Stack
The tools behind the work.
AI & LLM
- Bedrock · AgentCore
- SageMaker
- RAG & vector search
- Agent frameworks
- LLM evals
Cloud · AWS
- EKS · Lambda
- S3 · QuickSight
- IAM & networking
- Cost optimization
Platform · DevOps
- Kubernetes
- CI/CD pipelines
- Observability
- MLOps / deployment
Languages
- TypeScript
- Python
- Go
- Kotlin · Java
Depth on AWS specifically — not a flat list of every cloud. Tools are chosen to fit the problem, not the résumé.
Approach
How I approach engagements.
Most AI projects don't fail in the model — they fail in everything around it. These are the shapes of problems I'm usually brought in for.
From demo to production
Taking a RAG pipeline or agent that works in a notebook and making it reliable, observable, and affordable under real traffic.
Agents that don't go off the rails
Designing multi-agent systems with the guardrails, evals, and fallbacks needed to trust them with real workflows.
Keeping the cost curve sane
Architecting inference and infrastructure so spend grows with value, not just with usage.
Operating AI like software
Bringing DevOps/MLOps discipline so AI systems can be deployed, monitored, and rolled back like anything else in production.
Client work is kept confidential — no logos, no case studies, no metrics. What I can show you is how I think.
Read the full approachEngagement model
Ways to work together.
Independent and senior — you work directly with me. Pick the shape that fits where your AI work is right now.
Architecture & direction
Architecture reviews, technical direction, and de-risking an AI roadmap. Part-time and ongoing, for teams who mainly need senior judgement.
Embedded lead
Embedded as a senior engineer or architect for a set number of days a week — hands on the system, in the room for the hard calls.
Scoped delivery
A defined outcome: take a specific AI system from prototype to production, with clear scope and a definition of done.
Let's talk about your AI system.
If you're moving AI from prototype to production and want a senior pair of hands on the hard parts, I'd like to hear about it.