Description
We are looking for an Agentic AI Engineer to design, build, execute, test, and orchestrate autonomous AI agent systems that operate across complex, multi-step workflows. You will work at the intersection of large language models, tool-use frameworks, and enterprise data pipelines to deliver reliable, production-grade agentic solutions.
Responsibilities
- Design and implement agentic AI systems (single and multi-agent) with tool use, memory, and fallback mechanisms.
- Build production-grade agents using frameworks like LangGraph, AutoGen, CrewAI, or custom LLM orchestration layers.
- Implement agent reasoning loops including planning, tool selection, execution, observation, and re-planning with safety guardrails.
- Develop prompt and context engineering strategies for reliable, grounded LLM outputs.
- Design agent orchestration workflows include task routing, parallel execution, state management, retries, and human-in-the-loop escalation.
- Build evaluation frameworks for LLMs and agents including automated testing, adversarial testing, and performance benchmarking.
- Implement retrieval and grounding using vector databases, embeddings, and knowledge graphs for contextual accuracy.
- Ensure observability of agent systems by tracing LLM calls, tool usage, and decision paths using monitoring tools.
- Apply security and governance controls including prompt injection defense, access control, and protected tool execution.
- Optimize agent systems for latency, cost, and scalability in production environments.
- Build CI/CD pipelines for agent workflows including versioning, testing, and controlled deployments.
- Integrate agents with enterprise systems and APIs to automate end-to-end business workflows.
- Design feedback loops using production traces and evaluation signals to continuously improve agent performance.
- Experience with Model Context Protocol (MCP) systems to design database connections, integrate APIs, and enable secure tool orchestration for AI agent
📌 Agentic AI Engineer (Gurugram)
🏢 EXL
📍 Gurugram