AI Engineer — Multi‑Agent & Multi‑Modal Agentic Systems (Azure | Production Grade)
Location: Pan India
Experience: 4-8yrs
About the Role
We are hiring an AI Engineer with hands‑on experience building and operating production‑grade multi‑agent, multi‑modal AI systems on Azure. You will develop agent workflows that can reason, plan, use tools safely, and escalate to humans when needed while meeting enterprise requirements for reliability, security, observability, and cost/latency control.
Key Responsibilities
- Build multi‑agent orchestration (planner/router/specialists/validator) using LangGraph/LangChain (or equivalent) with explicit state and routing.
- Develop multi‑modal capabilities for documents + images (PDFs, scans, forms, screenshots) using OCR/document intelligence and vision understanding.
- Implement Multi‑Modal RAG: ingestion, chunking, metadata enrichment, indexing, retrieval, grounding, and structured/cited responses.
- Build tool‑using agents that integrate with enterprise systems (ticketing/service desk, internal APIs) with safe boundaries and auditability.
- Implement production controls: retries, timeouts, circuit breakers, rate limits, caching, token/cost budgets, and safe stop conditions.
- Establish evaluation + AgentOps: hallucination/grounding accuracy, tool‑call correctness, routing quality, containment/FCR, and structured error analysis.
- Instrument end‑to‑end observability (tracing spans across LLM calls, retrieval, tool calls, decisions, errors) and improve reliability over time.
Minimum Qualifications
- 4 - 8+ years of strong software engineering experience (Python preferred) with production delivery ownership.
- Proven experience shipping multi‑agent systems in production (LangGraph/LangChain or equivalent) with guardrails and failure handling.
- Strong experience with RAG (hybrid retrieval, ranking, metadata filters, evaluations).
- Experience delivering multi‑modal pipelines (document intelligence/OCR + vision; voice optional).
- Azure Cloud experience, including building and operating services using a stack such as:
- Azure OpenAI, Azure AI Search / Cognitive Search, Azure Functions or AKS, API Management, Key Vault, Entra ID, Application Insights.
- Strong API design, debugging, and distributed systems fundamentals.
Nice to Have
- ReAct / planner‑router‑verifier patterns; long‑running workflows, queues/state stores; event‑driven architectures.
- Microsoft Teams bot experience.
- Responsible AI / security guardrails for tool‑using agents; PII controls and audit logging.
