AI Engineer to Build Voice-Enabled Chat Assistant

We’re building a customer-facing AI assistant and need an engineer to take it from prototype to production. Users should be able to type or speak to the assistant, share images, and get accurate, safe responses. The system should run on our own servers. What you’ll do: Deploy an open-source language model as the core of the assistant Implement RAG so the assistant answers from our own documents and knowledge base Add stt so users can talk to the assistant, and tts so it can reply by voice Add image understanding so users can share photos or screenshots Build a moderation layer that: Blocks harmful or inappropriate content before it reaches the model Detects jailbreak attempts and users trying to trick the assistant into breaking its rules Keeps conversations on topic and refuses unrelated requests Checks the whole conversation, not just single messages, to catch attempts that build up over several turns Stays fast, so safety checks don’t slow down normal conversations Package everything so our team can deploy and maintain it Ideal candidate: Experience shipping LLM-powered products to production Hands-on experience with RAG, speech models and open-source LLMs Strong Python and API skills Pragmatic about latency, cost and reliability