About

I do the work of a five-person team. AI runs in parallel threads under my frameworks and evals, and twenty-five years of factory software tell me where to aim it.

  • Sole owner of the production test software, firmware programming infrastructure, and data backbone across three manufacturing sites — a team's scope, carried for years
  • Built for safety-critical controllers (SIL 3): production test, traceability, and the data systems behind them
  • The floor runs on it: data entered once, steps that cannot be skipped, reports that build themselves
  • Management gets answers in seconds; sensitive data never leaves the building
  • Two decades of SQL Server and .NET, widened by Claude, Gemini, and Codex under frameworks and evals I built
25+
Years in software & automation
3
Manufacturing sites supported
SIL 3
Safety-critical scope

Drive

I take over the stalled rollouts and finish them.

Integrate

I connect the systems you already run, so data is entered once and visible everywhere.

Automate + AI

I build the automation and AI that fill the gaps, on your hardware.

Approach

Scan in · track the floor · bridge the data

Scan at entry Tracked through WIP AI AI bridges the disconnects

Hours saved is the visible part of an AI project. Underneath sit the layers that decide whether it survives production. I map those layers first, then build.

Strategy

Audit the workflows, score each for AI readiness, model the return. Work starts where the measured value is highest — not where the change is easiest.

Data

An LLM call takes seconds; the pipeline behind it takes months. Most of the work is cleaning sources and filling gaps, and I start with the data you already have.

Infrastructure

Production needs more than a working demo. Pipelines, API layers, and the right model for each task keep the system reliable at scale.

The AI work

The part I'm known for: the context, guardrails, and evals that turn a model into production-grade output. If the eval score drops, it doesn't ship.

The human side

The system has to make the operator's day shorter. I design for adoption: fewer steps, clearer screens, visible wins.

Sovereignty

Everything can run on your own hardware. Sensitive data stays inside the building, which keeps your regulatory posture intact.

Longevity

I write it down. The system keeps running, and improving, when I'm not in the room.

Skills

Software & Architecture

System DesignData ArchitectureAPIs & Integration Internal Apps & ToolsCloud / CI-CDC# / .NET SQL ServerPython

Test & Automation

NI TestStandLabVIEWSTEM ATE / PXI / DAQHILDFT / DFM SPEATeradyne

Cyber & Safety

IEC 62443SIL 3 / IEC 61508Secure Boot PKIFirmware SigningCode Signing SDLThreat Modeling

AI Engineering

Context EngineeringMulti-model OrchestrationAI Guardrails RAG / Knowledge SystemsOn-prem LLMsAgentic Workflows Claude CodeGemini CLICodex
What's next

Let's connect your operation.

You are not adding an employee. You are adding the missing experience that lets your business ride the AI transformation without disrupting the workforce you already have. Open to architect-level roles and embedded engagements.

Get in touch