Job Description – AI Engineer Intern (Agentic AI)

About the Role

Key Responsibilities

1. Agentic AI & Agent Harness Development

• Design and develop AI agents using LangGraph, LangChain, and DeepAgents.
• Build and customize agent harnesses to support planning, tool execution, context management, and multi-step workflows.
• Implement agent middleware for tool retries, error handling, human-in-the-loop approvals, and context management.
• Work with agent memory, state management, checkpointing, and persistent execution.
• Develop reusable tools, sub-agents, and agent orchestration workflows.

2. Model Context Protocol (MCP) & Tool Integration

• Develop and integrate MCP servers and clients to connect AI agents with external systems.
• Build reusable tools for interacting with APIs, databases, and enterprise applications.
• Implement structured tool calling, input validation, error handling, and secure tool execution.
• Understand tool discovery, tool permissions, and integration patterns.

3. AI Observability & Monitoring

• Implement end-to-end tracing and monitoring for LLM and agent-based applications.
• Work with observability frameworks such as MLflow, LangSmith, Langfuse, or OpenTelemetry.
• Capture and analyze agent execution traces, tool calls, model responses, token consumption, latency, and failures.
• Debug agent execution paths and identify reliability and performance issues.
• Support monitoring and optimization of AI application quality and inference costs.

4. LLM Evaluations (Evals) & Reliability

• Design and implement evaluation frameworks for AI agents and LLM applications.
• Create evaluation datasets with ground-truth examples and expected outputs.
• Implement evaluations for tool selection, tool execution, structured outputs, and end-to-end agent decisions.
• Measure correctness, precision, recall, consistency, hallucinations, and task completion.
• Develop automated regression tests to evaluate changes in prompts, models, and agent workflows.
• Experiment with LLM-as-a-Judge and human feedback mechanisms.

5. Generative AI & Application Development

• Develop LLM-powered applications using Python and model APIs.
• Implement prompt engineering, structured outputs, function calling, and Retrieval-Augmented Generation (RAG).
• Build APIs and backend services using FastAPI or similar frameworks.
• Work with structured and unstructured enterprise data, including documents, emails, PDFs, and databases.
• Collaborate with engineering teams to develop and deploy AI solutions for real-world business use cases.

Required Technical Skills

• Programming: Strong Python fundamentals, OOP, asynchronous programming, REST APIs, and JSON.
• Agentic AI: Hands-on exposure to LangGraph, DeepAgents, or comparable agent frameworks, including agent orchestration and tool calling.
• Agent Harness: Understanding of agent execution loops, middleware, tool integration, state management, and error handling.
• MCP: Understanding of MCP architecture and experience building or integrating MCP servers and tools.
• Observability: Familiarity with instrumenting and debugging LLM or agent workflows using tracing and monitoring tools.
• Evals: Understanding of evaluation datasets, correctness metrics, and methods for evaluating LLM and agent outputs.
• LLMs: Knowledge of prompts, embeddings, structured outputs, function calling, and RAG.
• Software Engineering: Git, debugging, unit testing, and basic database knowledge.

Good to Have

• Experience with Azure OpenAI, OpenAI APIs, Anthropic, or open-source LLMs.
• Familiarity with MLflow, LangSmith, Langfuse, or OpenTelemetry.
• Experience with Pydantic, FastAPI, PostgreSQL, MongoDB, or vector databases.
• Knowledge of Docker, cloud deployment, and CI/CD pipelines.
• Exposure to multi-agent architectures, human-in-the-loop workflows, and persistent agent memory.
• Familiarity with document intelligence, information extraction, or enterprise workflow automation.

Educational Qualifications

• Currently pursuing or recently completed B.E./B.Tech/M.Tech/MCA in Computer Science, Artificial Intelligence, Data Science, Information Technology, or a related discipline.
• Strong programming fundamentals and demonstrated interest in AI engineering.
• Academic projects, open-source contributions, hackathons, or independent AI projects are an advantage.

Ideal Candidate

• Has explored building AI agents beyond basic prompt-based applications.
• Understands how agents interact with tools, manage state, and execute workflows.
• Thinks about AI reliability, evaluations, and observability.
• Can debug agent workflows and identify opportunities for improvement.
• Enjoys working on real-world engineering challenges.
• Is proactive, technically curious, and comfortable exploring rapidly evolving AI technologies.
• Can explain the architecture, implementation, and technical decisions behind their projects.
• Our ideal candidate is a hands-on AI builder who can design an agent, connect it to real tools, evaluate its decisions, observe its execution, and continuously improve its reliability.