AI Projects

Generative AI, RAG, Agentic AI, and Semantic Search Projects

Explore Urmila Saini's project work across RAG assistants, MCP-driven agentic workflows, semantic medicine search, FastAPI backends, vector databases, and Hugging Face Spaces deployments.

Saarthi AI project preview Live Space

Saarthi AI

Agentic Commute Intelligence Platform

Saarthi AI predicts commute risk using live traffic, weather, events, and route signals with an explainable departure recommendation.

  • Built an agentic platform that estimates lateness risk and recommends optimal departure times.
  • Integrated live traffic, weather, events, route-risk signals, and MongoDB Atlas state.
  • Shipped an explainable risk score with a why breakdown and deployed it on Hugging Face Spaces.
MCPMulti-AgentMongoDB AtlasLive APIsAgentic AI
SmartNotes AI project preview Static preview

SmartNotes AI

Handwritten Notes RAG Assistant

SmartNotes AI turns handwritten notes into a searchable study assistant using OCR, embeddings, retrieval, and AI summaries.

  • Built a RAG assistant that turns photographed handwritten notes into a searchable knowledge base.
  • Engineered chunking, indexing, and vector retrieval to keep responses grounded.
  • Enabled question answering, summarization, and study-material generation for handwritten content.
RAGOCREmbeddingsVector SearchLLM Apps
MediSwap project preview Live Space

MediSwap

Semantic Medicine Recommendation System

MediSwap finds composition-based medicine alternatives with fuzzy matching, semantic search, FastAPI, Qdrant, and embeddings.

  • Built a medicine recommendation platform for composition-based alternatives.
  • Combined fuzzy string matching with vector similarity over BGE embeddings.
  • Architected the retrieval layer with FastAPI, SQLite, RapidFuzz, and Qdrant.
FastAPIQdrantBGE EmbeddingsRapidFuzzSemantic Search