LLM Voice Assistant Backend

I’m building an LLM-powered voice assistant for smart glasses and I need a seasoned Python hand to take charge of the backend integration. The core services must handle real-time data processing from the glasses, route the streams through an LLM workflow, and expose reliable API endpoints the device can hit without latency hiccups. You’ll be working almost entirely in Python 3.11+, writing asynchronous, multithreaded code that can juggle audio chunks, context windows, and model responses in parallel. Everything lives in a GitHub repo, so disciplined branching and clear commits are a must. I already containerise the stack; your comfort level should be “intermediate” with Docker—think docker-compose files, multi-service networks, and tweaking images when we add a new dependency. Helpful extras: if you’ve deployed to AWS (Fargate, Lambda, or ECS), spun up MongoDB for session data, automated pipelines with GitHub Actions, or built REST / WebSocket routes in FastAPI, you’ll slot in even faster. Key deliverables • A fully-tested Python backend capable of real-time audio handling and LLM response generation • Clean, documented API endpoints for the smart-glasses firmware to consume • Docker configuration that stands up all services with one command and runs green in CI • Read-me level documentation so another dev can clone, compose-up, and reproduce results I review work in milestones: first a proof-of-concept service that echoes transcription in real time, then expansion into the full LLM pipeline, and finally optimisation and deployment. If you thrive on low-latency, event-driven code and enjoy seeing AI speak through wearable tech, let’s get started.

Project ID: 39730292