Source description
About the role
2+ years delivering enterprise-grade GenAI systems in production, including RAG pipelines, multi-agent and agentic workflows, function/tool calling, and MCP-based integrations with enterprise systems. Experience must extend beyond proofs of concept into live, business-critical deployments. 3+ years on a major cloud (AWS / Azure / GCP) and modern DevOps pipelines, including containers, infrastructure-as-code, and CI/CD, with hands-on ownership of deployment, observability, and cost and latency tuning for AI workloads. End-to-end delivery ownership, with a track record of working alongside architects and project managers to take AI initiatives from discovery through production rollout and steady-state operations. Strong grasp of GenAI evaluation, observability, and safety, with experience building evaluation harnesses, monitoring for hallucinations, tracking cost and latency, and implementing guardrails against prompt injection. Strong cross-functional collaboration, having worked alongside security, data engineering, and QA teams, and engaged directly with client SMEs through the delivery lifecycle. Team leadership of 3+ AI engineers, including mentoring, code reviews, and the ability to translate between business stakeholders and the engineering team. Must-have: Python, LLM APIs (OpenAI, Azure OpenAI, Anthropic, or Gemini), RAG architecture (chunking, embeddings, re-ranking), vector databases (Pinecone, Weaviate, pgvector, or Azure AI Search), agentic frameworks (LangChain, LangGraph, CrewAI, or AutoGen), MCP and function/tool calling, prompt engineering, LLM evaluation tooling (Ragas, LangSmith, or custom harnesses), at least one major cloud (AWS, Azure, or GCP), Docker, Kubernetes, CI/CD, infrastructure-as-code (Terraform), FastAPI and microservices. Good to have: Fine-tuning (LoRA or QLoRA), open-source LLMs (Llama, Mistral, or Qwen), voice and vision agents, GraphRAG and knowledge graphs (Neo4j).
More at Opteamix