Muhammad Hamza’s Official AI Engineering Portfolio
Client-to-production AI delivery, applied research, and software engineering case studies.
- All
- Computer Vision
- Agentic AI
- Generative AI
- Classical ML
- Final Year Project
- Data Science/Preprocessing
- Python | Django
- AI
- Cybersecurity
- Research
VisionGuard — Warehouse Safety AI
YOLO + ByteTrack computer vision system detecting forklift–pedestrian collision risk with real-world distance, TTC and explainable risk scoring.
AI / Computer Vision
NeuroInsight AI
Agentic Flutter clinical intelligence platform exploring 3D neuroimaging, explainable AI, biomarkers and case-aware RAG for Alzheimer's MRI analysis.
AI / Medical Research
Flood Prediction with LightGBM
End-to-end regression pipeline predicting flood-related outcomes from environmental data, with feature importance and full model evaluation.
Machine Learning / Kaggle
Attendance Management System
Cross-platform Flutter + Node.js/TypeScript workforce platform for attendance tracking, shift management and HR reporting.
Full-Stack / Mobile
AeroNav Sim — Drone Flight Simulation
Autonomous UAV simulation platform with A* path planning, PID flight control and multi-drone coordination, built toward ROS/PX4/Gazebo.
Robotics / Simulation
AI Virtual Painter
MediaPipe hand-tracking computer vision app turning a webcam into a gesture-controlled drawing canvas — no physical input device required.
Computer Vision
Aircraft Engine RUL Prediction AI
Predictive-maintenance ML system estimating turbofan engine Remaining Useful Life from NASA C-MAPSS sensor data (Random Forest, XGBoost, LightGBM).
Research Project
Agentic Cybersecurity AI Assistant
Multi-agent system + RAG to automate vulnerability triage, scheduled scans and PDF reporting for enterprise security operations.
AI / Cybersecurity
Agentic RAG-Powered Support Assistant
LLM-driven agentic assistant that autonomously resolves technical product queries with real-time, knowledge-grounded retrieval.
NLP / Automation
AI-Driven Intrusion Detection
Hybrid TabNet + Transformer deep-learning architecture for network intrusion detection, trained on the CIC-IDS 2017 dataset.
Cybersecurity / Research
3D-MobiBrainNet — Alzheimer's Detection
3D CNN multi-plane fusion model for multi-class Alzheimer’s classification (97.33% accuracy) — published in Ain Shams Engineering Journal (Elsevier).
Research | Publication
AI-Based Ticket Scalping Prevention
Anomaly detection and behavior analysis to detect bot purchasing and apply dynamic mitigation & pricing strategies.
AI / Fraud Detection
University Timetable System
Constraint-based automated scheduling web app for faculty and class allocation built with Django.
Django / Web App
LinkedIn Automation Bot
Selenium + BeautifulSoup based automation for profile scraping and connection workflows (data collection for outreach).
Web Automation / Data Collection
Satellite Image Classification
EuroSAT land-cover classification using a transfer-learned ResNet18, with confusion matrices, ROC curves and real-time inference.
Deep Learning / Research