Senior/Lead Python BE
Summary
Build and scale a Python-based AI avatar platform for a US client, integrating LLMs and handling high-concurrency video generation with FastAPI, Celery, and AWS.
Overhedge is looking for a strong Python developer for a long-term client project (US client).
About the project. An AI avatar platform: users create a video avatar. Heavy third-party integrations, high concurrency expected from launch. This is not an MVP: part of the codebase already exists and needs to be taken over, extended, and scaled for real traffic.
It is not a web3 project.
Stack:
Python, FastAPI (async code is a must)
PostgreSQL
Celery or similar queues (RabbitMQ, Redis) — long-running background jobs, video generation up to ~30 minutes
AWS, Docker
LiveKit / WebRTC for the video part (deep knowledge is a plus, not required)
Mandatory AI integrations: OpenAI and other LLM APIs, Langchain, RAG, tool calling, structured outputs.
What we expect:
5+ years of experience with Python as your primary language, not a secondary one
solid asyncio understanding: event loop, coroutines, what a blocking call inside a coroutine does
real experience with Celery or equivalents for long-running tasks
API resilience: retries and backoff, circuit breakers, webhooks (idempotency, out-of-order handling)
English B2+: you will communicate with the team and the client in English
production AI-integration experience is a strong plus
readiness to grow into the tech-lead role on this track is a plus
Terms:
full-time, fully remote
start: a paid ~2-week trial, then a 4-month contract moving to open-ended
process: a 15-minute intro call, then a technical interview (~1.5h, live coding plus discussion, no take-home task)