Hi, I'm Paramjit Singh

An intelligent document chatbot powered by Retrieval-Augmented Generation (RAG) that lets users upload PDFs, DOCX, or TXT files and interact with them through context-aware AI conversations. Built for fast semantic search, accurate retrieval, and multi-LLM responses using Groq and Gemini.

A deep learning-powered image captioning system that generates rich, human-like descriptions using both a pretrained transformer (Salesforce BLIP) and a custom CNN-LSTM architecture. Combines state-of-the-art inference with full pipeline model training for robust and flexible caption generation.

End-to-end ML-powered student risk prediction system. Processes 14k+ student records using a Random Forest + Logistic Regression ensemble to predict dropout risk levels. Features real-time predictions via Flask API, early warning prioritization, explainable AI insights, recommendation engine for mentors, and an interactive React dashboard with analytics, filtering, and CSV upload capabilities.

Applied K-Means clustering and PCA on DAIC-WOZ clinical depression database to discover 2 distinct depression subtypes from speech and text features. Achieved statistically significant results (χ²=6.44, p=0.0112) validating clusters against PHQ-8 clinical labels, demonstrating objective, data-driven depression diagnosis.

Intelligent sentiment analysis system for extracting actionable insights from 140K+ MOOC course reviews. Features FastAPI REST API, Streamlit dashboard, multi-model approach (LR, NB, RF, BERT), and 87% accuracy with real-time predictions.

Full-featured real-time chat application built with Node.js, Express, and Socket.IO. Features live messaging, multiple users support, file sharing, emoji reactions, private messaging, room management, and responsive design.

Interactive Data Structures and Algorithms visualizer built with React and modern JavaScript. Features algorithm animations, step-by-step execution, sorting visualizations, graph traversals, and educational interface.

Advanced Artificial Intelligence and Machine Learning system developed for the Indian Department of Posts. Uses Python, TensorFlow, and Data Science techniques to analyze demographic and economic data.

Advanced AI-powered text summarization application built with JavaScript and Google Gemini AI API. Features intelligent content analysis, automatic text summarization, and real-time AI processing.

Modern Next.js weather application built with React and API integration. Features real-time weather data, server-side rendering, location-based forecasts, responsive design, and optimized performance.

Professional Python game development project using Pygame library. Features advanced object-oriented programming, collision detection algorithms, dynamic scoring system, and progressive difficulty mechanics.