Work

What I've built.

I build side projects to use them, not to market them. Building is how I test what AI can actually do before I recommend it to a client.

This is what I've shipped. Email me if you want a heads-up when the next one is ready.

Nymara wordmark and logo

Nymara

Media library organizer for Jellyfin homelab owners

Live Windows Android

Nymara is a local-first media library organizer for Jellyfin homelab owners. It scans messy folders, previews a reorganization plan before moving anything, pulls TMDB metadata, scores duplicates by resolution, codec, and bitrate, and connects to the ARR stack. The AI runs on your own hardware through Ollama — no cloud required. Free, with a $25/year Pro tier.

I built it because I wanted a media organizer that kept my files local, never deleted anything by accident, and let me preview every change. Nymara taught me what local AI can do on a real Windows machine without enterprise infrastructure.

What it does for you

  • Search and discover media with plain language, with everything staying on your network
  • Let local AI agents handle complex library tasks in the Pro tier
  • One portable .exe for Windows and one app for Android, with Velopack auto-updates
  • Plugs into Jellyfin plus the ARR stack you already use
  • Preview and undo every reorganization before it touches your files
Ollama Semantic Kernel .NET Android Velopack
Visit Nymara

Behind the build

The hard part was not pulling metadata — it was making a destructive file reorganization safe. Every move gets a dry run, an undo tree, and protected folders. Building Nymara proved that local AI on a real Windows machine can match cloud convenience for this workflow.

Nymara mascot

Next project

In progress — details when it ships

I have a few more side projects in various states of done. If you want an early look, send me a note and I’ll tell you when the next one is ready.

Want something built?

Same approach on your project.

I work the same way I build: understand the real problem, choose the right tool, document what I did, and leave you with something that works.

See AI Consulting Services Email Me

I reply within 2 business days. Or email me: [email protected].

What I build with

Tools and experience.

Sixteen years of explaining tech to people who don't have time for jargon. That matters more than any specific tool. But here are the ones I go deep enough in to actually ship with:

Local AI

Ollama, Semantic Kernel, local model orchestration, and agents that run on your own hardware instead of someone else's cloud.

Integrations

Jellyfin and the ARR stack, Proton Mail, Stripe, Resend, Retell, VoIP routing — making separate tools talk to each other so you don't have to.

Stacks I ship in

Python + FastAPI, .NET + WPF, PHP + Apache, JavaScript/TypeScript, Astro, Cloudflare Workers. I pick the tool that fits the job, not the one that's trending.

Infrastructure

Cloudflare, Vultr VPS, Linux, DNS, email deliverability, and the unglamorous work that keeps a site or service online.