Now

Leading automation strategy across enterprise platforms - designing scalable Playwright framework patterns that can be reused across client projects, and building out the AI-assisted engineering workflows the team uses for test authoring, triage, and review. The interesting problem this month is getting agent-driven workflows reliable enough to trust with production-grade quality gates.

Going deeper on agent architectures - particularly multi-step orchestration on AWS Bedrock and tool-use patterns with Claude. Also catching up on the latest in autonomous QA research: how do you actually verify the behaviour of an AI-driven workflow when the workflow itself is non-deterministic?

This portfolio site, which doubles as a sandbox for things I want to try (an interactive terminal hero, a serverless contact form with inquiry routing, a handful of automation-themed mini-games).

Not actively job-hunting - settled in my current role. Not on Twitter / X. Not taking on any new long-term contracting commitments. The contact form is still open for short consulting conversations, mentorship requests, and anyone with a genuinely interesting automation problem.