Job description
About the role
As an AI Engineering Intern, you'll work directly with our core engineering team to design, build, and evaluate AI-powered systems that solve real problems at scale. This is a hands-on, high-autonomy role with meaningful responsibility from day one.
What you'll do
Design and develop AI agents and coding agents using leading LLMs, integrating tool use, memory, and multi-step reasoning. Build and optimize Retrieval-Augmented Generation (RAG) pipelines — from chunking and embeddings to vector stores and reranking. Implement Model Context Protocol (MCP) integrations and custom tool harnesses to extend LLM capabilities into real-world systems.
Craft and systematically refine prompts (zero-shot, few-shot, chain-of-thought), documenting what works and why. Build AI workflows that orchestrate multiple models, tools, and data sources into end-to-end applications. Construct evaluation (eval) frameworks to benchmark model accuracy, reliability, and regression. Instrument AI systems with observability tooling — traces, token usage, latency, failure modes, and cost tracking.
Contribute to generative AI applications across text, code, and multimodal contexts. Technical skills Large Language Models (LLMs)
- Generative AI
- AI Agents
- Coding Agents
- Prompt Engineering
- Retrieval-Augmented Generation (RAG)
- Model Context Protocol (MCP)
- Tool Harnesses
- AI Workflows
- Evaluations & Benchmarking
- Observability
- Python Who you are Currently pursuing or recently completed a B.Tech / M.Tech / MS / PhD in Computer Science, AI/ML, or a related field. Candidates from premier engineering institutions (IITs, IISc, NITs, IIITs, BITS Pilani, or equivalent) are encouraged to apply. Strong Python programming skills, with clean, readable, production-aware code. Genuine curiosity about how LLMs work, and a habit of experimenting independently. Familiar with at least one LLM orchestration framework (Lang. Chain, Llama. Index, CrewAI, Haystack, or similar). Exposure to vector databases (Pinecone, Weaviate, Chroma, pgvector) and embedding models is a plus. Strong written communication — able to document findings and share learnings clearly. Work policy This role is full-time at our Hyderabad office. We do not offer remote or hybrid arrangements; candidates must be willing to work on-site and relocate to Hyderabad if required.