Gaurav Singh

Data Scientist (AI Product Team) at Newgen Software

Developed a knowledge-base RAG system with contextual chunking, Gemini embeddings, and Maximum Marginal Relevance (MMR) search, plus a Parent Document Retriever to deliver contextual answers. Enriched document metadata with summaries and keywords using an LLM, improving agent performance by 20%. Implemented a verification layer with the G-eval library and used opik for evaluation management. Architected and deployed a Jira agent on GCP leveraging Vertex AI Hybrid Search and LangChain; built modular, stateful workflows to fetch and generalize latest tickets, storing them in a vector database for quick, context-aware answers with inline citations; reduced internal ticket search time by 50%. Designed a POC AI agent for an insurance client to support underwriting decisions and claims processing; multi-agent architecture with LangGraph and RAG with hybrid search; integrated rule-based validators with LLM-based reasoning to ensure regulatory compliance and reduce manual review by 35%.

Data Scientist Intern at HRS Group

Designed a cross-team analytics dashboard to visualize 6-week and 12-month trailing timelines, increasing decision-making speed through better data visibility. Created ETL pipelines using AWS Glue, PySpark, and S3 to automate data processing and transformation tasks.