Services
Framed as outcomes, because that's what you're buying. Every engagement uses the same methodology: interview-driven planning, guardrails as standard, evidence before assertions, staged cutovers.
An embedded engineer, for a defined outcome
The forward deployed model: I sit inside your business for the length of one outcome, working with the people who own the problem. Discovery with your operators, a hardened plan, production code in your environment against your real data and constraints, and accountability until it runs. Not a slide deck and a handshake; a system, live.
An AI workflow running in production
We pick one business process (a backlog, a reporting cycle, a nightly operation) and I build the agentic workflow that runs it: wired into your existing tools, with human checkpoints where they matter and a staged cutover that never puts live operations at risk. You end with a system in production, not a proof of concept.
A data platform that runs itself
Microsoft-stack data engineering (Fabric, Azure, Power BI) with the agentic layer on top: pipelines that observe themselves, report their own nights in plain English, and grow a troubleshooting knowledge base with every incident. Built from scratch or retrofitted onto what you have.
Your team, enabled
The practices installed so they outlive the engagement: proven procedures packaged as reusable skills, planning and review workflows your team runs themselves, and the safety patterns that let you trust automation with real systems. The goal is that you stop needing me.
Hiring instead?
The same portfolio serves both conversations. I'm open to senior AI engineering and forward deployed engineer roles where production agentic systems, customer-facing delivery, and platform ownership are the job. The case studies are my CV in long form; the short form is available on request.