Himanshi Sharma

I’m an AI Engineer focused on building practical Generative AI applications—especially Retrieval-Augmented Generation (RAG) systems and AI workflow pipelines powered by LLMs. I’m drawn to work where reliability matters, from prompt engineering to structured output generation and evaluation.

In my projects, I developed a Fact-Checking RAG Assistant using LangChain, Gemini API, Tavily Search, and Hugging Face. I implemented vector retrieval, response evaluation workflows, and LLM-as-a-judge validation to reduce hallucinations and improve output consistency.

I also built a Conflict Resolution Multi-Agent System with LangGraph and LangChain. It includes memory, tool calling, and structured task routing, along with modular state-based workflow pipelines that support contextual multi-agent reasoning and automation.

Most recently, at Fulcrum Digital (Consultant Intern, AI Engineering), I worked on an AI-driven insurance claim processing system using LLMs, OCR, and FastAPI. I developed document intelligence and data extraction pipelines, built rule-based validation workflows to reduce processing errors, and supported API integration, prompt engineering, workflow testing, and evaluation pipelines for AI-generated output.