GenAI Engineer

We are looking for a Senior GenAI Engineer to join a high-impact AI initiative focused on automating complex, multi-step document generation workflows through state-of-the-art AI agents. The platform transforms raw clinical and unstructured data into structured, compliant outputs at scale. This is a hands-on builder role. You will code every day, working within an existing agentic architecture to extend, optimize, and improve the GenAI-facing components of the platform. What You'll Do Build and optimize multi-agent orchestration systems and autonomous workflows Design and scale advanced RAG pipelines over large, heterogeneous datasets Implement state management, short/long-term memory, and self-correction loops within agent networks Create evaluation metrics, observability pipelines, and guardrails for hallucination reduction and content quality Optimize AI inference cost, latency, token usage, and system reliability Implement security controls including access management, data encryption, and prompt injection mitigation Document and share reusable agent patterns and prompt libraries across engineering teams Required Skills GenAI & Agentic Frameworks LangChain, LangGraph, LlamaIndex, CrewAI, AutoGen Advanced RAG, GraphRAG, prompt engineering, fine-tuning, function/tool-calling Schema-constrained outputs, semantic caching Vector Databases At least one of: FAISS, Pinecone, Qdrant, Milvus, Weaviate, Pgvector LLM Infra & Serving vLLM, Hugging Face TGI, or Ollama Guardrails & Validation Pydantic, Instructor, NeMo Guardrails or Guardrails AI Observability Ragas, DeepEval, Langfuse, or LangSmith Engineering Python (FastAPI, asyncio), Docker, Kubernetes, CI/CD, Git, OpenTelemetry MCP and Agent2Agent protocols a plus AWS (Preferred) Amazon Bedrock, SageMaker, OpenSearch, Step Functions Nice to Have Experience in pharma, biotech, or life sciences environments Familiarity with clinical document workflows (protocols, study reports) Knowledge of GxP or 21 CFR Part 11 compliance frameworks Exposure to CDISC standards (SDTM, ADaM) Profile 6+ years in software/data engineering, including 2-3 years hands-on GenAI/LLM in production Comfortable working autonomously within an existing technical architecture Strong communicator : able to document complex systems and collaborate with cross-functional teams Show more Show less