AI Engineering
Multi-agent orchestration under engineering controls. I build the platforms that make engineers faster — then measure whether they actually did.
↺ evidence, not vibes
Containerised, security-hardened development environments; AST-based code intelligence with incremental, content-hash-keyed indexing; hybrid semantic search; and multi-agent pipelines that plan, build, review and iterate under real engineering controls.
Two things, deliberately separate. Hands-on training that takes engineering teams from AI-assisted development to full multi-agent orchestration, run against their own repositories — and the installable that makes it stick: a hardened agent sandbox and a codebase-semantics layer committed into their own repos.
Teams whose customers are other engineers: development loops, CI/CD, build and release. Happy engineers, faster iteration, safer releases — measured, not asserted.
Every productivity claim is a hypothesis. I A/B benchmark agent tooling on real production tickets against human ground truth — measuring cost, correctness and coverage — and I have shipped, redesigned and killed features on that evidence, including publishing results that contradicted my own initial claims.
That discipline is the product. Anyone can demo an agent; the work is knowing which parts survive contact with a real codebase, and being willing to say so when they don't.
ProcuraconCurrent
First engineering hire
Inherited an existing codebase and accelerated delivery by building an AI-forward development platform around it: hardened containerised environments, AST code intelligence and multi-agent delivery pipelines.
OrchLabCurrent
Principal Advisor
AI engineering consultancy moving teams up the agentic maturity curve — from chat and copilot, through spec-driven development, to Level 4 orchestration where engineers coordinate specialised agents to plan, design, build, test and deploy at scale.
DriveSaverCurrent
Founder
Consumer app, built and shipped end to end through my custom agentic framework.
Snap
Engineering Manager
Led the Content Creators Foundation team — a team whose customers were three other engineering teams.
Amazon
Engineering Manager
Led engineering teams on the Prime Video worldwide checkout funnel, and on Fire TV across BBC iPlayer and Amazon TVs.
INX Software
Engineering & Engineering Leadership
Development loops, build and release for a product engineering team.
Two decades of engineering, from 12-person startups to Amazon and Snap — mostly running teams whose customers were other engineers.
Perth, Western Australia — working remotely with distributed teams. Reach me directly about speaking or collaboration. For consulting and training engagements, start at orchlab.ai.