Job title: Agentic AI Trainer Intern
Location: Vijayawada
Duration: 3–6 months (extendable)
Stipend: Competitive
Short summary
We’re seeking an Agentic AI Trainer Intern to help design, train, evaluate, and document autonomous AI agents and agentic workflows using LLMs, RAG, and orchestration frameworks. This role is ideal for final-year students or recent grads with hands-on Python skills and interest in LLMs, prompt engineering, and evaluation metrics.
Key responsibilities
- Build, test, and iterate on agent workflows and multi-step LLM agents using frameworks like LangChain, LangGraph, or similar.
- Create and maintain prompts, few-shot examples, and instruction templates for agent behaviors and tool use.
- Develop and run evaluation tests (task success, hallucination rate, latency, and reliability) and report improvements.
- Prepare training datasets, design synthetic task scenarios, and label / curate data for retrieval-augmented generation (RAG) pipelines.
- Integrate agents with APIs, simple tools (e.g., calculators, web search, file readers), and vector DBs (e.g., Pinecone, Milvus, Weaviate).
- Help maintain experiment logs, reproducible pipelines, and documentation for handoffs to engineering teams.
- Assist with user-testing sessions, collect feedback, and iterate on agent behavior and safety/guardrails.
Must-have qualifications
- Pursuing or recently completed B.E./B.Tech/MCA/M.Sc in CS, IT, AI/ML, Data Science, or related field.
- Strong Python skills and familiarity with Git/version control.
- Working knowledge of LLMs, prompt engineering, and RAG concepts.
- Basic understanding of vector embeddings and vector databases.
- Good analytical, communication, and documentation skills.
Work Location: In person
