Key Responsibilities
Design and build agentic LLM solutions (single- and multi-agent patterns) to solve real business problems across domains (e.g., customer support, document intelligence, knowledge retrieval).
Build RAG pipelines end-to-end: data ingestion → chunking/embeddings → vector search → retrieval orchestration → response synthesis, with measurable quality.
Implement prompt engineering and prompt orchestration (prompt chains, tool-calling, function calling), including prompt iteration and cost/latency optimisation.
Develop production services/APIs for LLM applications (e.g., FastAPI/Flask/Streamlit) and integrate with enterprise systems and data sources.
Apply guardrails to reduce hallucinations, enforce policy constraints, and ensure secure tool usage; implement evaluation strategies for LLM and RAG outputs.
Collaborate with Data Engineering teams to ensure data quality, governance, and documentation standards, and with MLOps/Platform teams for CI/CD, monitoring, and reliable deployments.
Create and maintain technical documentation, solution design artefacts,
and reusable components for faster delivery and consistent engineering practices.
Must-Have Skills
5 to 12 years total experience, with hands-on LLM/GenAI delivery experience (preferably 1–3+ years building production-grade LLM apps).
LLM / GenAI & Agentic Engineering
Hands-on experience with LLMs including Claude (Anthropic) and other leading models; robust understanding of capabilities, limitations, and use-case fit.
Practical experience with RAG, embeddings, vector databases (e.g., FAISS/Pinecone/ChromaDB), semantic search, and retrieval quality evaluation.
Experience with frameworks/tools such as LangChain, LangGraph, Hugging Face, or equivalent orchestration stacks.
Experience building agentic workflows including tool calling/function calling; familiarity with “agentic architecture” concepts is valued.
Exposure to Claude Code or similar coding-agent workflows is a plus (
📌 Agentic/ai Engineers With Claude/code/llm Ski Noida
🏢 EXL
📍 Noida