What I work on

Six areas where I build, and where I have opinions worth arguing about. Each one is a primer on what is genuinely hard, not a summary of what the field already says.

UNS at the edge

UNS, MQTT, OPC-UA, S7, NATS, Kafka, i3X and MCP — and the data contracts that make any of it mean something.

Fleet management

k3s fleets across a hundred sites: GitOps, offline-first reconciliation, commissioning, and handing the keys over.

Observability

Performance-controlled OpenTelemetry, tagset governance, metric triage, and the traces industrial systems never have.

Planning optimization

Planning and scheduling, constraint modelling, and optimisation.

Models under a latency budget

ML engineering for performance: profiling, quantisation.

GPUs on the line

Accelerated inference in production: drivers, CUDA and TensorRT, Triton serving, and GPU sharing on edge nodes.