13 years running multi-site federal logistics taught me what breaks, what scales, and what operators actually need. Every tool in this portfolio was built to solve a problem I lived. The ops background is what makes the AI useful.
I build AI-powered operational systems and deploy them. The 13 years running federal logistics is what makes them solve real problems instead of demo ones.
Every tool in this portfolio was conceived, architected, and shipped on personal time — not as coursework, not as speculative concepts, but as working software built to scratch a real operational itch. GOATvision aggregates 16 live data endpoints into a single situational brief. GOATflow ingests raw operational chaos and extracts prioritized action. Melvin is a voice-first AI companion with per-user memory isolation. These are not side experiments.
The right opportunity is one where the ability to think in systems, build proof of concepts, and translate AI capability into operational reality is the job — not a nice-to-have.
Michael J. Hinkle III
AI Builder & Operations Leader · Seattle, WA
Each project below was independently conceived, architected, and shipped while leading full-time multi-site operations. This is not a portfolio of side experiments. It is a demonstration of applied AI at the practitioner level.
Aggregates 16 live data endpoints across 9 distinct APIs — traffic, transit, ferries, weather, fire dispatch, seismic activity, and flight delays — into a single static HTML file with no login, no backend, and no build step. A rule-based Ops Brief synthesizes every panel into a plain-English situational summary of the metro area in real time.
Expanded to DMV and Philadelphia as live secondary deployments with feature parity. Some data sources — Pennsylvania 511, live traffic cams, PulsePoint emergency dispatch for DMV — require commercial API access or LLC-gated credentials. Those requests are in flight. The architecture is already built to accept them.
Each city runs on a shared Supabase telemetry layer with location-key separation and 90-day rolling data retention. Built to scale to 4-5 cities on a single free-tier instance.
Python + Streamlit, powered by Google Gemini 1.5 Flash. Ingests raw operational input — emails, documents, voice memos, images — extracts prioritized action items called Tracks, and drives execution through a gamified dual-currency reward system and 7-level ascension engine.
A voice-first AI companion built on Firebase auth, ElevenLabs voice integration, and per-user memory isolation. Features a gated invite system and persistent context across sessions. The most architecturally complex build in the portfolio — and the one I'm most focused on right now.
The ecosystem that GOATflow, GOATvision, and Melvin all feed into. WorkGOAT is a B2B workforce performance platform that credentials individual work ethic over time — turning small wins into a verifiable behavioral record. Fresh Cheese earned in GOATflow ports to WorkGOAT at 1:1 on launch.
Routes users to the most effective AI tool for their specific task — reducing cognitive friction in tool selection and improving AI workflow efficiency through intelligent decision routing. Built in React, TypeScript, and Tailwind with Gemini Flash.
An AI-powered executive briefing workflow that synthesizes daily operational data into a structured, actionable morning brief. Built on Gumloop with a multi-node automation architecture. Named in honor of my mother, Debra.
Compresses carrier onboarding from weeks to days using AI-guided route sequencing, real-time field coaching, and adaptive performance benchmarking. Published as a formal Concept Overview and Commercial Value Assessment. The operational problem is real — I managed it firsthand across multiple sites.
Interactive workforce education platform covering the full arc of AI — from the 1950s to now. History, working tools, infrastructure fundamentals, energy, robotics, medicine, quantum computing, and forward-looking trends. Built for operators and managers, not engineers.
Includes 8 free tools from OpenAI, Google, Amazon, and Microsoft with hands-on access and guided learning paths. Deployed independently as a practical resource for teams navigating the AI transition.
Earned formal recognition across multiple assignments — not for tenure, but for measurable operational outcomes. These awards represent district-level acknowledgment of turnaround performance in environments where most managers request transfers.
From managing 10,000+ service addresses across distributed geographies to personally covering routes during a December blizzard with 75% of the carrier workforce unavailable — the work has always been grounded in a simple principle: the mission doesn't pause because conditions are difficult.
I lead senior operations and program management at the federal level while building AI systems independently. I'm selectively evaluating opportunities where field-tested execution and hands-on AI development are both valued. If that describes your organization, let's talk.
