Hello, I'm
I love to Code AI with AI
Transforming complex AI concepts into practical solutions. Specializing in NLP, Computer Vision, and Cloud AI Engineering.
I'm Muqadas Ejaz, a passionate AI & ML Engineer focused on building reliable, user‑centric intelligent systems.
I specialize in model development, experimentation, and deploying AI/ML systems to the cloud. I enjoy turning complex problems into simple, scalable solutions.
PDF QA RAG System is an AI-powered application built with LangChain, LLaMA, and Streamlit, designed to provide accurate question-answering from PDF documents. The system leverages LangChain’s retrieval-augmented generation (RAG) pipeline and LLaMA’s LLM capabilities to extract precise answers directly from the document. Streamlit enables an intuitive and interactive user interface, allowing users to upload PDFs, ask natural language questions, and receive reliable responses—with the system responding “I don’t know” if relevant information is not found.
Human Detection System is a real-time person detection system built with the YOLOv8 deep learning model. The system processes input from either a webcam (live detection) or a video file (offline detection), enabling flexible usage. It accurately identifies all persons in each frame, draws bounding boxes with labels, and counts the total number of persons detected, making it suitable for applications in surveillance, crowd analysis, and safety monitoring.
MuseMate 🎨🤖 is a creative and conversational AI chatbot built with Streamlit and LLaMA 3.1-8B-Instruct from Hugging Face. It features a clean, interactive web interface that enables users to engage in natural, dynamic conversations with an open-source LLM. Designed for creative writing, casual chats, and idea brainstorming, MuseMate provides a friendly and playful AI companion that blends functionality with style.
A machine learning application using Logistic Regression to predict cardiovascular disease from clinical features. Achieved 88% accuracy and deployed with a Flask-based GUI for real-time predictions.
Micro Facial Expression Recognition is a deep learning project focused on detecting subtle, involuntary facial expressions using Convolutional Neural Networks (CNNs). These micro-expressions occur briefly and often reveal genuine emotions, making them highly valuable in fields like security, psychology, and suspect interrogation. The model is trained and evaluated on two benchmark datasets: CASME II (micro-expressions) and FER-2013 (facial expressions). By combining these datasets, the system enhances its ability to distinguish both subtle and broader emotional cues, contributing to more accurate and reliable emotion detection.
Heart Disease Prediction System is a machine learning-powered application designed to predict the likelihood of cardiovascular disease based on key clinical features. Early and accurate prediction supports better clinical decision-making and improved patient outcomes. Developed a Logistic Regression model trained on the Cleveland Heart Disease dataset. Evaluated performance using accuracy, precision, recall, F1-score, and confusion matrix, achieving an 88% accuracy rate. Built a Flask-based GUI that allows users to input patient data and receive real-time predictions. Ensured the codebase is clean, modular, and reproducible, making it suitable for research, education, or real-world prototyping.