Duane Gomes · Enterprise & AI Architecture
Designing enterprise AI systems that are agentic by design and governed by default — with the observability and control planes that make them safe to run in production.
Project Portfolio
Working systems, each on its own subdomain with source open for review. Built to demonstrate agentic patterns and responsible-AI operations end to end.
Live demo
A multi-agent enterprise-architecture mesh: specialised agents that collaborate across capability maps, roadmaps and portfolios — showing how agentic orchestration replaces brittle point-to-point automation.
Live demo
An AI control-plane reference: tracing, evaluation and cost/quality telemetry for LLM workloads — the instrumentation that turns a demo into something you can responsibly operate.
In progress
Additional reference builds are in development across agentic governance and evaluation. New tiles will appear here as each demo goes live on its own subdomain.
Thought Leadership
A practitioner series on operating AI responsibly at enterprise scale — observability, control planes and the governance that production AI demands. Published on Medium.
Why observability is the precondition — not the afterthought — for running generative AI in production, and what to instrument first.
Read on Medium →The architecture that sits between models and the enterprise — routing, policy, evaluation and cost governance as a first-class control plane.
Read on Medium →The next instalment in the series is in draft. Follow on Medium to catch it when it publishes.
Follow the series →of technology experience — architecture, strategy & governance
Let's talk
I advise and architect enterprise AI — agentic systems, responsible-AI governance, and the platforms that make them production-ready. If that's the conversation you're having, let's have it.
