Open to opportunities

Gaurav Singh

@gauravsingh15

Message

Data Scientist building agentic RAG and AI systems that improve accuracy, compliance, and operational speed.

India

Message

What I'm looking for

I’m looking to build production AI—especially agentic RAG, NLP, and hybrid search on cloud—where I can improve accuracy with strong evaluation/verification, collaborate across product teams, and ship measurable impact on real workflows.

I’m a Data Scientist focused on turning complex information into reliable, context-aware AI experiences. At Newgen Software, I developed a Knowledge Base agentic RAG system that used contextual chunking, Gemini embeddings, and MMR search, improving agent performance by 20%.

I also architected and deployed a JIRA agent on GCP for internal sales and support teams. By using Vertex AI Hybrid Search and LangChain with modular, stateful workflows, I reduced ticket search time by 50% through quick, cited, context-aware answers.

For an insurance client, I designed a POC decisioning support system with a multi-agent architecture using LangGraph. The solution combined RAG with hybrid retrieval and added rule-based validation agents to support compliance, reducing manual review time by 35%.

Earlier, I boosted Q&A retrieval accuracy by 25% within an internal LangChain-based RAG product, built NLP pipelines for sentiment/intent from call recordings with a 30% processing-time reduction, and delivered consulting work including a 92.5% accurate credit card cross-sell model. I bring rigorous evaluation practices and an engineering mindset—so results are measurable, factual, and production-ready.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Newgen Software

Aug 2023 - Present (2 years 10 months)

Developed a knowledge-base agentic RAG system using contextual chunking, Gemini embeddings, and MMR with a Parent Document Retriever, improving agent performance by 20% with a G-eval verification layer and opik evaluations. Architected and deployed a GCP Jira agent using Vertex AI Hybrid Search and LangChain, reducing ticket search time by 50%, and built an insurance decisioning POC with multi-age

Agentic RAG Gemini Embeddings MMR Retrieval Lang Chain LangGraph Vector Databases G Eval Hallucination Verification Opik Evaluations

Data Scientist Intern

HRS Group

Apr 2023 - Jun 2023 (2 months)

Built cross-team analytics dashboards to visualize 6-week and 12-month trailing trends, improving decision-making speed by 20%. Created ETL pipelines using AWS Glue, PySpark, and S3 to automate data processing and transformation tasks.