Narendranath Athota

AI/GenAI Developer with hands-on enterprise experience at CGI, building Agentic RAG systems and multi-agent LLM applications.

Designed

LangGraph-based agentic workflows and hybrid retrieval pipelines that improved retrieval quality by 40%, reduced token usage to 2K–2.5K per query, and enhanced response reliability. Proficient in FastAPI, RAGAS evaluation, prompt engineering, and end-to-end Generative AI development.

Apprentice AI Developer - CGI - Bengaluru, India

(2025-09)

• Built a modular enterprise Agentic RAG platform using LangGraph, ChromaDB, and tool-calling agents for adaptive retrieval and grounded response generation across 5+ production use cases.
• Optimised context management and retrieval strategies to reduce token usage to 2K–2.5K tokens per query, lowering overall token consumption by 30–40% while improving response accuracy and cost efficiency.
• Developed a hybrid search and retrieval pipeline using BM25, FAISS, and reranking strategies, improving retrieval quality by 40% compared to baseline approaches.
• Automated a RAG evaluation workflow using LangChain and RAGAS across 100+ enterprise documents, evaluating relevancy, faithfulness, and retrieval quality while reducing manual review cycles by 40%.
• Implemented ReAct-style agentic workflows with self-evaluation and retry mechanisms, improving response consistency and reliability for ambiguous and complex queries.
• Supported production monitoring, API testing, troubleshooting, and operational coordination while collaborating with cross-functional teams and client stakeholders.

B.Tech - Computer Science (Artificial Intelligence) - KKR & KSR Institute of Technology and Sciences (2025-05)