Real-Time AI Orchestration Engineer (WebRTC / Mobile) - PT/FT - Worldwide

Role Overview

Founder-led stealth consumer AI product. We're building a real-time, voice-first AI experience (mobile, iOS + Android). The core engineering challenge — and the reason this role exists — is orchestration: we run expensive, GPU-backed live streaming sessions that are billed per minute with no native pause or resume. Most of the time the user doesn't need the live session; when they do, it has to feel instant. Your job is to architect that — start and stop those sessions on demand, switch between streaming vendors invisibly under a strict latency budget, and hide the spin-up — while keeping cost-per-minute under control.

Required (non-negotiable):

• Genuine, production experience with real-time streaming media — WebRTC, or frameworks built on it (LiveKit, Daily, Pipecat).
• Built a live, low-latency conversational voice pipeline — real turn-taking, barge-in, sub-second response.
• React Native, working competency with React Native, a very big plus if you have experience shipping apps to the App Store / Google Play
• Thinks in cost and latency at the streaming layer (cost-per-user-per-minute).
• A hands-on individual builder who can point to real shipped work.

Not a fit if your "real-time voice" is a managed platform (Vapi, Retell, Dialogflow) or plain WebSockets — those aren't the media layer we need.

Engagement: Remote contract, ~30+ hrs/week, ~3 months to build the prototype, potential to extend. Must overlap with US Eastern.