Capabilities, in context.

A skills page is easy to inflate, so this one is deliberately anchored: every capability below maps to systems described in the case studies - trading engines held to p99 commitments, claims platforms under audit, and a 51-agent AI system run in production. Levels mean what they say: expert is where I have been paged at night and fixed it; working knowledge is where I ship carefully with the docs open.

The through-line across all of it is systems that make decisions under load: event-driven backends on Java, Go and Node, Kubernetes and AWS underneath, Postgres and queue-first designs in the middle, and AI agent orchestration on top. The maths section is not decoration - queueing theory, probability and optimisation are what turn "add more workers" into a capacity plan. Each group below links back to the case study where that skill earned its level, so you can judge the claim against the work rather than against a progress bar.