Most developers add AI. I build systems that make your product smarter than your competitors'.
Not ChatGPT wrappers. Not tutorial projects dressed up as experience.
I'm talking about RAG pipelines with pgvector, voice interfaces with ElevenLabs, on-device inference with ONNX Runtime for users with zero connectivity, and 3D visualization with Three.js all in production.
At my last role, I led a team of 5-10 developers building a field service platform for technicians who work in dead zones no internet, no excuses.
The hard part wasn't the AI. It was architecting a system that worked online with OpenAI + RAG, and switched seamlessly to on-device ONNX models when signal dropped. Then adding voice commands so technicians could complete forms hands-free while under a car.
That's the kind of problem I like. The ones most freelancers won't touch.
A few things I've shipped you can look at tonight:
HealthMind AI — Full-stack health companion with symptom analysis, mood tracking, and AI insights. Built in TypeScript with Prisma. 104 commits. Not abandoned maintained.
CryptoSage — Real-time crypto news with sentiment analysis and AI-generated signals. Built for traders who need signal, not noise.
SerenAI — Voice-guided meditation using ElevenLabs conversational AI. Because sometimes the interface should disappear entirely.
Tattoo AI Studio — AI generates the design, Three.js previews it on a 3D body model before the needle moves. Clients love it. Artists close faster.
Here's what actually separates me:
Most freelancers start at the API call. I start at the data layer embeddings, retrieval architecture, inference strategy then build up. By the time we hit the UI, the intelligence is already baked in.
I've done this across health, finance, wellness, and creative industries. The stack adapts. The approach doesn't.
If you're building something that needs to actually think not just respond I'd like to hear about it.
