The RAG architecture follows this process:
User question → Query routing → Relevant information retrieval → Context generation → LLM response → Source information
Doom AI: can answer questions about my:
Built a Bengali/English RAG system for Bangladesh VAT law QA with page-level citations. Implemented hybrid dense +
BM25 retrieval, Reciprocal Rank Fusion, and cross-encoder reranking for reliable legal document retrieval.
AI-Powered Agricultural Decision System — a full-stack social + AI platform for farmers and agricultural communities.
Social media features: post, like, comment, add friends
Crop disease detection using Computer Vision (YOLO)
AI-powered analysis & recommendations via LLM integration
Paste any YouTube link. Our AI breaks it into quizzes, flashcards, and a day-by-day study calendar — so you actually remember what you watched.
An autonomous agricultural rover that combines robotics with AI for smart farming field operations.
Dual mode: fully autonomous and remote control operation
Automatic crop disease detection using onboard cameras
Automated soil sample collection from infected areas
A RAG-based intelligent chatbot for United International University that answers queries using live university data.
Web crawler for scraping and indexing UIU website content,
Vector database (Pinecone) for semantic search & retrieval,
Groq-powered LLM for fast, accurate response generation,
Retrieval-Augmented Generation (RAG) architecture
