class UqashaZahid:
"""AI/ML Engineer | Electrical Engineering Student @ UET Lahore"""
def __init__(self):
self.name = "Uqasha Zahid"
self.role = "AI / Machine Learning Engineer"
self.education = "B.Sc. Electrical Engineering (AI) — UET Lahore (2022–2026)"
self.location = "Lahore, Pakistan 🇵🇰"
self.contact = "uqashazahid@gmail.com"
self.core_domains = [
"🔍 Computer Vision & Image Intelligence",
"📈 Anomaly Detection & Pattern Recognition",
"⏱️ Time Series Forecasting & Analysis",
"⚡ Energy AI & Smart Grid Security",
"🧑💻 Real-Time AI Inference & Deployment"
]
self.ai_skills = [
"Deep Learning", "Machine Learning",
"Neural Networks", "Transfer Learning",
"Object Detection", "Behavioral Analysis"
]
self.frameworks = ["TensorFlow", "PyTorch", "Keras", "Scikit-learn"]
self.deployment = ["Streamlit", "Hugging Face Spaces", "Docker", "Azure AI"]
self.languages = ["Python (Advanced)", "C (Embedded)", "MATLAB/Simulink"]
self.certifications = [
"✅ Microsoft Certified: Azure AI Engineer Associate (2025)",
"✅ NAVTTC: Advanced ML & Data Mining (2024)"
]
def current_focus(self):
return "⚡ Dual-Stage Deep Learning SCADA System for Electricity Theft Detection"
def fun_fact(self):
return "My AI scans 40,000+ consumers & 423 transformers with 70% less compute! 🔋"
🔋 A production-grade AI-powered SCADA system for detecting electricity theft across the Lahore power grid. Built with a dual-stage deep learning engine, live GIS mapping, and a cyberpunk-themed interactive dashboard.
| 🔧 Component | 📋 Description | 📊 Scale |
|---|---|---|
| 🧠 Stage 1 — Grid Monitor | Deep learning model for transformer-level anomaly detection | 423 Transformers |
| 🔬 Stage 2 — Behavioral Scanner | Neural network for deep consumer usage pattern analysis | 40,000+ Profiles |
| ⚙️ Smart Cost Optimizer | Skips verified-safe sectors to save computing resources | 70% Less Compute |
| 🗺️ Live GIS Mapping | Real-time geographic visualization across the power grid | Interactive |
| 👥 Multi-Role Dashboard | Dedicated panels for Utility Admins and Consumers | Streamlit |
| 🐳 Cloud Deployment | Fully containerized and live on Hugging Face Spaces | ✅ Live |
╔══════════════════════════════════════════════════════════╗
║ ⚡ PowerGuard AI — How It Works ║
╠══════════════════════════════════════════════════════════╣
║ ║
║ 📡 Grid Data ──► 🧠 AI Stage 1 ──► Anomaly? ║
║ (Deep Learning Model) │ ║
║ │ YES ▼ ║
║ NO ✅ Safe 🔬 AI Stage 2 ║
║ (Skip Deep (Neural Behavioral ║
║ Scan) Analysis) ║
║ │ ║
║ ▼ ║
║ 🚨 Theft Detected! ║
║ 📊 Auto Report Generated ║
╚══════════════════════════════════════════════════════════╝
Production System |
Deep LearningComputer VisionBiometric AIAutomation
| ✨ Feature | 📝 Details |
|---|---|
| 👤 Biometric Recognition | High-dimensional facial embeddings with real-time identity tracking |
| 👥 Production Scale | Fully deployed and operational for 100+ users |
| 📱 Smart Alerts | Automated absence notifications via WhatsApp integration |
| 🖥️ Native Desktop App | Windows UI with role-based access — Student & Admin panels |
| 📊 Live Monitoring | Real-time attendance logs and profile management dashboard |
AI Motion Control |
Computer VisionDeep LearningPose EstimationReal-Time AI
| ✨ Feature | 📝 Details |
|---|---|
| 🦴 Pose Estimation | Full-body 33-point skeletal landmark detection using AI |
| ⚡ Ultra-Low Latency | Sub-100ms end-to-end real-time inference pipeline |
| 🎮 Virtual Controller | AI translates body gestures into keyboard and joystick inputs |
| 🕹️ Gaming Experience | Seamless, touchless character control via motion intelligence |
Smart Surveillance AI |
Object DetectionMulti-Object TrackingComputer Vision
| ✨ Feature | 📝 Details |
|---|---|
| 🔴 Violation Detection | Red-light jumping detected via AI-based signal state classification |
| 🔁 Persistent Tracking | Robust vehicle identity retention even during visual occlusions |
| 📄 Automated Evidence | Auto-generated timestamped violation reports with incident data |
| 🛣️ Multi-Lane Coverage | Simultaneous AI monitoring across all lanes in real-time |
╔══════════════════════════════════════════════════════════════╗
║ 🧠 Competency Overview ║
╠══════════════════════════════════════════════════════════════╣
║ Python & AI Scripting ████████████████████ 95% ║
║ Machine Learning ████████████████████ 95% ║
║ Computer Vision ███████████████████░ 90% ║
║ Deep Learning & Neural Nets ███████████████████░ 90% ║
║ Data Engineering ██████████████████░░ 85% ║
║ AI Model Deployment ████████████████░░░░ 80% ║
║ Cloud AI (Azure) ██████████████░░░░░░ 70% ║
║ Embedded & Systems (C) ██████████████░░░░░░ 70% ║
╚══════════════════════════════════════════════════════════════╝
| 🏅 Certification | 🏢 Issuing Body | 📅 Year | 🎯 Domain |
|---|---|---|---|
| Azure AI Engineer Associate | Microsoft | 2025 | ☁️ Cloud & AI Engineering |
| Advanced Machine Learning & Data Mining | NAVTTC | 2024 | 🤖 Machine Learning |
| Final Year Project — AI SCADA System | UET Lahore | 2026 | ⚡ AI & Energy Systems |
National Centre of Artificial Intelligence (NCAI), UET Lahore | Jun 2025 – Sep 2025
| 🔧 Responsibility | 📋 Details |
|---|---|
| 🤝 Research Collaboration | Worked alongside Senior Research Engineers on applied AI projects |
| 🏥 Healthcare AI | Developed machine learning models for medical classification and prediction |
| 🔒 Security AI | Designed and optimized real-time computer vision pipelines for security systems |
| 📁 Data Engineering | Processed and annotated 2,000+ medical images including normalization and augmentation |
| 📈 Performance Optimization | Applied transfer learning to significantly improve model accuracy and detection metrics |
| 🎓 Degree | 🏛️ Institution | 📅 Duration | 🎯 Specialization |
|---|---|---|---|
| B.Sc. Electrical Engineering | University of Engineering & Technology (UET), Lahore | 2022 – 2026 | Artificial Intelligence |
| 🚀 Project | 🔧 Technology | 🏷️ Domain | 🔗 Access |
|---|---|---|---|
| PowerGuard AI — Electricity Theft Detection | Docker · Streamlit · Deep Learning | AI · Energy · Smart Grid | ▶ Open Live App |
