Presentation
adopted AI coding agents and shipped LLM features, and cannot yet prove the output is safe, observable or cost-controlled.
I have taken agentic development from experiment to production standard inside a live multi-repo estate — authoring the organisation's AI development standards,
building the agent orchestration tooling, and shipping an LLM-backed incident pipeline — while owning the security, privacy and release boundaries around it.
Underneath that sits nineteen years of Node.js, TypeScript and AWS in regulated and high-throughput domains: banking (HSBC), crypto compliance (Xapo), online
trading (Interactive Investor) and betting (OpenBet). I do the architecture and I stay in the code.
Available for fractional retainers, fixed-scope audits and project delivery.
