I’m building a production-ready Voice AI module and need a full-stack partner who can take it end-to-end. On the client side everything runs in Next.js; on the server side I’m using Python with FastAPI. Your job is to wire the two together, weave in modern voice-recognition capabilities, and leave the whole stack fast, reliable, and easy to scale. Here’s what I expect: • Frontend – Ship clean, reusable React / Next.js components that handle microphone access, status feedback, and whatever UX you propose for speech capture and playback. SSR should remain intact and Lighthouse scores can’t drop. • Backend – Stand up new FastAPI endpoints that receive audio streams, call the recognition logic, and return results with low latency. The code must be fully typed, documented, and covered by basic tests. • Voice layer – Plug in best-fit speech-to-text and text-to-speech engines (open-source or cloud, I’m flexible) so we can recognise spoken phrases and synthesise responses. Accuracy and cost efficiency are both important. • Integration – Make frontend calls seamless through WebSocket or REST, whichever yields the smoothest real-time experience. CORS, auth tokens, and error handling need to be airtight. • Performance – Benchmark round-trip time, add caching or queueing where it helps, and keep memory/CPU footprints sensible so we can containerise and auto-scale later. Deliverables are a Git repo with modular code, a one-page setup guide (Docker-compose preferred), and a short Loom walkthrough showing the feature in action. If you’ve shipped something similar in Next.js + FastAPI, tell me. Otherwise, outline your plan, any libraries you’d lean on, and how long each milestone would take. *Share your previous project