Uday Bhan Singh

Experience

1. Data Scientist (Uber Gss Uberai) - Uber Ext

   Indium

   Dec 2023 — Present

   GenAI-Powered Multi-Agent Analytical Chatbot for Instant Business Insights: Engineered enterprise scale multi-agent analytics chatbot using Google ADK, with in-memory metadata retrieval, schema knowledge graph, multi-layer NL2SQL planning with sufficiency scoring, Python auto-code generation, and auto-context compression to deliver low-latency, 40% cache hit, 97% precise data-driven insights from structured datasets, automating complex analysis through a single message. Tech stack: Google-adk, neo4j, jinja2, gemini/gpt, FastAPI, Azure devops, App service, Github action3. Knowledge Management Enterprise RAG System: Built enterprise RAG system for Uber\'s knowledge management with semantic-aware parsing, adaptive chunking, post-processing deduplication pipeline, rich metadata engineering (structural/semantic/temporal/custom), hybrid retrieval (α=0.5), cross-encoder reranking, relationship graph mapping, comprehensive evaluation (Precision@5: 0.94, Recall@10: 0.91, MRR: 0.88, NDCG@10: 0.89, faithfulness: 0.92, coherence: 0.91), deployed on K8s with 99.95% SLA, handling 50K+ QPS3. Structured Metadata Generation for Documents using Llama3 and FAISS with Langchain Implemented an automated metadata generation system using Llama2 with RAG techniques and FAISS on AWS, dramatically reducing manual efforts by enabling efficient retrieval and contextual metadata creation. Streamlined workflows with a user-friendly Streamlit interface, significantly boosting project scalability and data manageability.

Education

• Motilal Nehru National Institute Of Technology

  Bachelor of Technology, Electrical and Electronics Engineering

  2015 — 2019