I build AI systems for operations teams in regulated industries, and I lead the teams that ship them. Pharma, health-tech, e-commerce, and contact centers.
Now: engineering and AI delivery through FRI Solutions. Production agent systems, retrieval pipelines, and self-hosted open-source infrastructure.
Before: Director of Software Engineering at Waterfield Technologies, where I grew the group from 5 to 40 engineers and owned Xcelerate, an AWS platform handling over 10 million customer interactions a month for enterprise contact centers.
clawops: MCP-native CLI for deploying and operating self-hosted OpenClaw instances across AWS, GCP, Azure and bare VMs. Read-only mode, destructive-action confirmation, and audit logging built in, because letting an agent touch production infrastructure needs a permission model, not good intentions. 493 tests, ADRs, schema generation enforced in CI, published to npm.
arcGIS-Rust-runtime: Rust runtime for ArcGIS workloads.
Most of what I build is client work and stays private:
- a component-sourcing agent that cut specialized fulfillment time by more than 85%
- a clinical-trial data pipeline with PII redaction feeding Snowflake
- a LoRA-tuned open-weight deployment serving over a billion tokens a month at roughly a tenth of frontier API cost
- generative AI use case delivery across vaccine franchises at Merck Happy to walk through any of it.