I turn real problems into shipped products — AI video pipelines, intelligent desktop apps, full-stack dashboards, and mobile software that people can actually use.
I produce broadcast video for Metro Nashville by day. Nights and weekends, I build the AI tools I wish existed for media work.
The pattern is simple: notice friction, build the thing, ship it.
Work that has to serve an operator, survive a deadline, or earn a place on someone’s device — not just look convincing in a demo.
Long-form sermon video in. Ready-to-post vertical clips out. A production pipeline chains scene detection, YOLO subject tracking, Whisper transcription, Gemini moment selection, reframing, captions, and export.
Research-grade validation, live execution, risk coordination, and monitoring for automated prediction-market strategies.
A private, single-user health system that keeps food, weight, habits, check-ins, and coaching inside one calm daily loop — no account required.
Drag in a clip, get a useful filename. Electron sends keyframes through a Cloudflare Worker to GPT-4o-mini Vision. No account, API key, or setup for the end user.
A native SwiftUI reading app that generates grade-tuned children's stories, with StoreKit 2 subscriptions and a Cloudflare Workers backend.
A real client-facing web presence for an automotive paint and equipment business — clear service storytelling, conversion-focused layout, and production-ready frontend work.
A local-first footage library that turns a mapped archive into searchable, previewable b-roll — built for editors who need the right shot quickly.
I learn by building at the edge of my current ability — then I keep going until the thing works for someone besides me.
By day, I produce broadcast video for Metro Nashville. That gives me a useful bias: I notice friction, I care about the operator, and I know the difference between an impressive prototype and a tool that survives contact with a deadline.
That mindset shows up in everything I make — APIs with failure modes, UIs with a clear next action, and systems that can run unattended without pretending the hard parts don't exist.
A practical stack shaped by the products above — from model integration to shipped clients and production services.
Python · TypeScript · JavaScript · Swift · SQL · Shell
React · Next.js · Vite · Tailwind · SwiftUI · Electron
FastAPI · Node.js · Cloudflare Workers · Docker · Vercel
OpenAI · Gemini · YOLO · PySceneDetect · Whisper · PyTorch
pandas · Chart.js · real-time dashboards · backtesting · validation
GitHub · CI/CD · App Store deployment · Tailscale · macOS automation
Open to remote software, AI engineering, and product-minded builder roles. If the problem is real and the team cares about the details, I want to hear about it.