Senior AI Agent Engineer — RAG, LLM Orchestration & Production Voice/SMS Agents

We're an established custom software studio shipping production-grade AI agents for real businesses — receptionists, schedulers, recruiters, and executive assistants that handle live calls and messaging at volume. We live in the agentic stack: RAG pipelines, tool-calling, multi-agent orchestration, and tight telephony integration. You'll work alongside a team that knows the difference between a slick demo and a system that survives contact with real production traffic. We're looking for an engineer to help extend and harden our agent platform. Depending on fit, the work may include: Designing and tuning RAG pipelines — chunking strategy, embeddings, hybrid search, re-ranking — over pgvector (or Qdrant/Pinecone) Building agentic workflows with tool-calling and structured outputs (LangGraph, LlamaIndex, or custom orchestration) Wiring LLMs into real-time voice and SMS via Twilio, including latency-sensitive turn-taking Writing evals so we measure quality instead of guessing at it Driving down hallucination and improving grounding and citation accuracy You should have: Real production RAG experience — not a tutorial chatbot. You can talk about retrieval quality, eval metrics, and why your chunking choices mattered. Strong command of modern LLM tooling: function/tool calling, embeddings, vector DBs, prompt and context engineering Node.js and/or Python; PostgreSQL a plus Bonus: voice agents, Twilio, MCP, fine-tuning/distillation To apply: skip the generic pitch. Tell me about one RAG or agent system you actually shipped — what broke, how you measured quality, and what you'd change. A short Loom or a code link beats a long cover letter.