We're building an agentic AI platform that orchestrates intelligent automation across enterprise systems.This role owns LLM reasoning pipelines end-to-end — from prompt design to production deployment on Azure Kubernetes.
What You'll Do
- Build LangGraph DAG workflows with deterministic state management
- Design LLM reasoning pipelines with structured JSON output and validation layers
- Integrate Azure OpenAI (GPT-4o) with confidence scoring and anti-hallucination guardrails
- Convert unstructured documents into structured process maps
- Implement observability (Langfuse) and checkpointing (CosmosDB) for agent state recovery
- Collaborate on architecture, prompt design, and compliance guardrails
Must-Have
- 4+ years Python building production services
- Hands-on experience with LLM APIs (OpenAI / Anthropic / Azure OpenAI)
- Strong prompt engineering and structured output design
- Experience with LangChain / LangGraph / CrewAI / AutoGen
- Pydantic or JSON schema validation
- Docker, Kubernetes, and CI/CD experience
- Familiarity with NoSQL databases
Strong Plus
- LangGraph supervisor patterns
- LLM observability (Langfuse / Helicone)
- Azure stack: AKS, CosmosDB, Azure OpenAI
- Process mining or rule-based validation systems
- Experience in compliance-sensitive environments
- Claude Code , Cursor as tooling
