Rahul Pandey

AI Engineer with a backend engineer's instincts and an AI practitioner's toolkit. Over 2.7+ years of experience in designing and shipping production-grade LLM integrated applications at TCS. Hands-on with RAG architectures, hybrid retrieval (TF-IDF + dense embeddings, RRF), and LangChain-based orchestration. Seeking a role where I can own AI products from architecture to deployment.

Software Developer at Tata Consultancy Services (TCS) (2023-08 – Present)

• Designed and integrated an intelligent document parsing pipeline using Python and LLM-based extraction to auto-classify insurance policy documents and route them into downstream processing queues — reducing manual triage effort by ~60% across high-volume claim ingestion workflows.
• Engineered the document ingestion and embedding pipeline for the claims chatbot parsed 500+ workflow SOPs and policy documents into semantically chunked FAISS embeddings with metadata filtering by claim type, reducing irrelevant retrieval results by ~40%.
• Integrated LangChain orchestration to chain retrieval, context injection, and LLM response generation for the claims assistant, cutting average query resolution time by 35% compared to manual documentation lookup.
• Developed the Product within a microservices architecture serving 30,000+ insurance entity records; resolved N+1 query issues and introduced composite indexes, reducing p95 API latency from 620ms to 380ms (39% improvement).

Associate Software Engineer Intern at Keross R&D (2023-01 – 2023-05)

• Built responsive Angular UI components connected to backend REST APIs; fixed 20+ cross-browser bugs and reduced customer-reported UI issues by 30%.
• Delivered UI screens across 5+ product pages aligned to design specs, reducing QA revision rounds within sprint cycles.

B.Tech in Computer Science & Engineering – IEM, Kolkata (2019 – 2023)