Solo dev with an audio background. Agent systems, real-time audio, and the engineering underneath.
Twelve years in entertainment and audio production first, then increasingly serious on solo engineering. Now building agent orchestration, audio engines, and the supporting infrastructure. Solo, AI-augmented, from Eindhoven (NL).
Work I can show
ECHO
My own agent orchestrator. Local, voice-first, with an Obsidian vault as memory. Multi-brain routing, tool-dispatch, three-tier AI fallback. Daily use.
AudioLaunch
Mobile music studio + Resonara (gamified sound design). Custom Rust audio engine for low-latency DSP on Android. Eight years solo-built, now in public open beta.
AudioLab.tools
Authority platform at the intersection of audio + AI + DSP + Android + accessibility. Real tools, deep docs. Astro + React + Tailwind v4.
TherapyVault
Secure video-sharing platform for therapists ↔ clients. Granular access-grants, GDPR data-export, MFA. Next.js 15 + Prisma + Cloudflare R2.
CaptionCompass
Accessibility app for deaf and hard-of-hearing Android users. Live captions + direction-of-arrival (only when reliable). Kotlin + Compose + GCC-PHAT.
PhotoFlow
From edited photo to online in 30 seconds. Lightroom plugin + Tauri desktop + multi-agent publishing pipeline for photographers.
FamilyKart
Mario Kart for my kids. Own 3D engine on top of Three.js, AI opponents, powerups, chase camera. Electron desktop app.
A reusable launch approach, crystallised while building and launching AudioLaunch. Tools, sequencing, and a custom content pipeline. Being applied to the next project.
What's under NDA I'm happy to walk through on a call.
Patterns
Engineering decisions I hit along the way. Short writeups, ~3-5 min each.
Recent writing
The most recent work I've written up in public — short writeups from real projects, not from a content calendar.
- Multi-brain LLM routing
Cheap heuristics first, an LLM classifier as fallback. Picking the right model tier per request without burning cost or quality.
- Three-tier AI fallback
Claude → rules → hardcoded. A degradation pattern that keeps the product working when the API doesn't.
- Real-time audio in Rust
Lock-free ring buffers, allocation-free audio paths, FFI patterns. Why Rust survives the real-time deadline.
What I bring
Strange combo: over ten years of audio DNA and someone who builds and uses agent systems daily. For audio products I'm in both worlds at once. For pure AI engineering I'm fluent with the tooling. No ramp-up time. No team. Little drama.
Frequently asked questions
- What does Nathan Renting do?
- I'm a solo developer from Eindhoven who builds AI-agent systems (LLM orchestration, tool-calling, RAG) and real-time audio engines — from prototype to production.
- Is Nathan available for freelance / contract work?
- Yes. I'm available for contract engineering (ZZP), 2-3 days/week, EU remote. I'm based in Eindhoven and reachable throughout NL for a kickoff or critical review.
- Which technologies does Nathan use?
- For AI agents I reach for Python, FastAPI, Claude API, MCP, LangGraph and RAG. For real-time audio: Rust, Kotlin, Android NDK and DSP. For product: Next.js, TypeScript and React Native.
- What does it cost to hire Nathan?
- Indicative €110-130 per hour. I keep project pricing negotiable depending on scope and duration.
- Does Nathan work remote or on-site?
- I prefer remote within the EU. On-site in the Netherlands works for a kickoff or a critical review.
- What makes Nathan's profile unique?
- I combine 10+ years of audio and DSP background with building agent systems daily. For audio-AI products I'm in both worlds at once.
Open for freelance / contract work (ZZP), 2-3 days/week. Remote-preferred from Eindhoven, reachable throughout NL for a kickoff or a critical review.
→ How to hire me