Agentic AI – Software Intern (B)
Job Category: Internship
Job Type: Full Time
Job Location: India - Gurgaon India - Kolkata Pakistan - Islamabad
Build and orchestrate AI agents to automate workflows and power intelligent tools. You’ll use LLMs, LangGraph, prompt engineering, and FastAPI to design multi-step, multi-agent systems that interact with data, APIs, and users.
Responsibilities
- Design, implement, and evaluate agentic workflows (single and multi-agent) using LLMs and LangGraph.
- Build FastAPI services that expose agent capabilities and tools.
- Engineer prompts, system instructions, and guardrails for reliability and safety.
- Integrate agents with internal data sources, vector stores, and 3rd-party APIs.
- Instrument experiments; track quality with offline evals and live metrics.
- Write clean, tested Python; document architectures and decisions.
Minimum Qualifications
- Strong Python programming skills.
- Working knowledge of AI/ML fundamentals: supervised/unsupervised learning, NLP, GenAI concepts.
- Experience with at least one LLM framework/workflow tool (e.g., LangGraph, LangChain) or willingness to ramp fast.
- Familiarity with REST APIs and FastAPI (or similar) for backend services.
- Clear communication and a bias to ship.
Education
BS/MS students in CS/EE/DS or related.
Preferred
- Hands-on with multi-agent environments (task routing, tool use, memory, evaluation).
- Retrieval and vector databases (FAISS, Pinecone, pgvector, etc.).
- Basic Docker use for local dev and packaging.
- Java familiarity is a plus.
Tools You’ll Touch
Python, LangGraph, OpenAI/other LLM APIs, FastAPI, Docker, Git, basic CI, telemetry/eval tooling.