I build the systems that run hyperscale data center delivery, and I build them by directing AI coding agents.
I'm a Sr. Power Platform & AI Developer at Microsoft (Cloud Operations + Innovation), building the cost-reporting, controls, and automation layer for the global data center capital program. Behind the tooling are 11+ years of delivering physical infrastructure:
- Managed a 6-person Turner & Townsend consulting team embedded with Google Data Centers' equipment supply chain, on a 1+ GW hyperscale data center portfolio
- Did Primavera P6 schedule forensics for construction disputes at Ankura
- Delivered a $200M NYC transit hub on schedule and $17M under forecast
M.S. Construction Management, Columbia · B.S. Mechanical Engineering, UVM
| Project | What it is |
|---|---|
| GPU Price Index · live dashboard | An open, daily index of GPU rental prices across hyperscalers, neoclouds, and marketplaces. It is built from primary sources, versioned in git, and collected in CI, with a cluster cost calculator and an MCP server so agents can query it. |
| MacroGauge · data center page | Daily data center cost indexes for build, operations, and hardware (switchgear, transformers, generators, HVAC, labor), plus state cost parity and real data center construction spending. Every published artifact is schema-validated and QA-gated, with an append-only revision history. |
| Sector Research | An agentic research analyst. It turns SEC filings, fundamentals, and earnings calls into cited theses with falsifiable kill criteria, then monitors them. The math is deterministic code; the model handles judgment, and its output is checked by an eval harness, 1,100+ tests, and CI gates. |
| Multi-Agent Market Research + Equity DD Agent Playbook | Fifteen parallel data and synthesis agents with deterministic guardrails behind a 42-tool MCP server, plus the due-diligence methodology written as executable agent skills. |
- Agent-directed and terminal-first. I write the spec, agents write the code, and I review, test, and evaluate it. I use Claude Code and Codex for independent work and GitHub Copilot CLI at work.
- Deterministic where it has to be right. Math, routing, and validation live in code, and the model handles judgment. Evals and CI gates check the model's work.
- Data products that keep running. They publish on a schedule, keep history, and fail loudly.
Stack: Python · SQL · FastAPI · Next.js · Postgres · GitHub Actions · MCP · Anthropic / OpenAI / xAI APIs · Power Platform (Power Apps, Power Automate, Power BI, Dataverse) · Azure · Primavera P6
I'm interested in roles where data center and AI-infrastructure delivery meets software and AI:
- Data center delivery TPM or controls lead roles that build the tracking, reporting, and tooling layer, not just coordinate it
- Applied AI and internal-tools roles inside data center, capacity, and infrastructure organizations
- Deployment lead or strategist roles taking AI systems from contract to production
I'm based in Florida and remote-first, and open to relocating for the right role. You can reach me on LinkedIn.