AI engineer focused on voice AI, agentic systems, and full-stack product work. I like taking ideas from first principles through production — no shelf-ware.
-
vowel — Open-source real-time conversational voice platform. Drop-in voice AI for any web app, no replatforming required. Custom voice models that compete with Google and ElevenLabs at a fraction of the cost. TypeScript, React, Cloudflare Workers, self-hosted and SaaS paths.
-
amebic — React-based graphics library for generating image assets from composable React components. Works for humans in a live preview UI or AI agents rendering assets programmatically. Deterministic by design.
- 🎙️ Voice AI and conversational agent tooling
- 🛠️ Open-source developer tools with an AI angle
- 📡 WebRTC and real-time communication systems
- 🤖 Agentic workflows and AI infrastructure
- Languages: TypeScript, Go, Python
- Frontend: React, Next.js, Vite, Three.js / React Three Fiber
- Backend: Node.js, Convex, Cloudflare Workers
- Infra: Kubernetes, Docker, Talos Linux, Terraform, Ansible, OCI
- AI / ML: Custom voice models, vLLM, llama.cpp, Vercel AI SDK, OpenAI and Gemini Realtime APIs
8+ years shipping software end to end. Recently focused on voice AI and agentic systems — building the infrastructure, the models, and the product layer. Previously: WebRTC streaming, distributed systems, 3D web tooling, and e-commerce platforms.