Implemented a Vision Transformer inspired by the paper 'An Image is Worth 16x16 Words', achieving high accuracy on food image datasets through patch tokenization and transformer-based classification.

Developed an AI-powered intelligent networking assistant that analyzes and summarizes LinkedIn and Twitter profiles to automatically generate actionable professional insights using LangChain and OpenAI API.

Fine-tuned a TinyLlama-1.1B model using LoRA to build a personalized and efficient command-line assistant optimized for limited compute resources and reduced training costs.

A Retrieval-Augmented Generation (RAG) system that ingests Medium blog articles and enables question-answering using vector similarity search. Features document processing, semantic search, and intelligent retrieval for context-aware responses.

Turn any public website into a bite-sized summary with the power of LLMs. Features a clean, chat-style Gradio interface that lets you drop a link and get AI-generated summaries. Uses Groq LLM API for high-quality summarization.

Built a Retrieval-Augmented Generation (RAG) system that processes nutrition textbooks and answers queries using semantic search and local Gemma language models. Implemented end-to-end pipeline with PDF processing, embedding generation, and GPU-optimized inference with 4-bit quantization for memory efficiency.