Staff Technical Program Manager working at the seam where program leadership meets engineering — turning ambiguous, cross-functional problems into systems teams can actually operate. My through-line is metal to cloud through real-time AI orchestration: carrier-grade infrastructure early on, AWS-native data platforms in the middle, real-time AI orchestration today.
- 🔭 Currently: Staff TPM at Intuit on the Virtual Expert Platform — the AI-driven layer behind expert help — leading real-time customer-to-expert routing across chat, phone, and email and shipping custom AI agents
- 🛠️ Focus: privacy, Trust & Safety, and AI governance — turning regulatory and policy signal into systems teams can run
- 🌱 Recently built: a real-time AI governance pipeline (below)
- 📫 Reach me: LinkedIn · mfrinzell@me.com
📌 Featured project — realtime-ai-governance
Metal to cloud through real-time AI orchestration.
A real-time pipeline that ingests a public privacy/regulatory feed (FTC, EUR-Lex), routes each item through a multi-agent classify → assess → register flow, and emits a versioned governance register that a daily job keeps current — a diffable, auditable record of what surfaced and when.
- Deterministic, zero-API-key core — reproducible in CI, runs anywhere
- Optional, explicitly gated LLM classifier
- Automated tests, lint, and CI; a scheduled job that commits an auditable artifact
A working demonstration of the pattern, built to extend — more sources, alerting, a dashboard — rather than claimed as a production system.
- Program leadership: large-scale carve-outs and migrations, real-time data and AI platforms, cross-functional delivery across engineering and operations (PMP, active)
- Privacy / Trust & Safety: led a regulator-driven account-integrity program under an FTC consent order — risk separation, signal-based detection, and the documentation behind it
- Technical fluency: Python, CI/CD, GitHub Actions, system and pipeline design
- Range: built a PMO from scratch through a 20,000-customer / 1,800-partner infrastructure carve-out; established experimentation and data-platform programs at consumer scale