class RanaMuhammadAhmed:
def __init__(self):
self.name = "Rana Muhammad Ahmed"
self.university = "Bahria University Islamabad"
self.cgpa = 3.9
self.location = "Rawalpindi, Pakistan 🇵🇰"
self.background = "Pre-Medical → Computer Science"
@property
def currently(self):
return [
"🌿 Researching edge AI for agriculture — CropCop (EfficientNet · TFLite · Flutter)",
"🛡️ Building AI security pipelines",
"📱 Shipping cross-platform Flutter apps in production",
"⚔️ Competing at ICPC Asia Regional level",
"🌍 GlobalX Student Ambassador @ Bahria University",
]|
Bahria University Islamabad | 2024 – Expected Dec 2027
|
Kips College, Rawalpindi | 2021 – 2023
|
| Domain | Tools |
|---|---|
| Languages | C++, Python, Dart, JavaScript, HTML5, CSS3, SQL |
| Mobile / Cross-Platform | Flutter, Dart, GoRouter, Flutter Quill, Supabase |
| AI / ML / Edge | PyTorch,Hugging Face, TFLite (INT8), FastAPI, Microsoft Presidio, Ollama / TinyLlama, spaCy, NLP |
| Systems & Embedded | Qt Widgets, SFML, CMake, Arduino, TinkerCAD, Multithreading |
| Developer Tooling | Git, GitHub, VS Code, Qt Creator, Android Studio, Kaggle |
| Core CS | OOP, Data Structures & Algorithms, System Architecture, Memory Management |
🌱 CropCrop: Edge AI Crop Disease Intelligence | PyTorch · Flutter · TFLite · EfficientNet | Active Research
An end-to-end self-supervised + multi-task learning pipeline for diagnosing plant diseases, pest damage, and nutrient deficiencies directly on low-end Android hardware, no internet required. Edge-deployed and offline-first.
| Accuracy | Top-1: 90.41% | Macro F1: 90.39% | Top-5: 99.43% |
| Performance | GPU Inference: 16.8ms |
| Model Size | 48MB (FP32) | 12MB (INT8) |
| Architecture | EfficientNet-Lite4 Backbone |
| Resources | Notebook, datasets, and backbone models available on Kaggle |
🛡️ LLM Security Gateway | Python · FastAPI · Microsoft Presidio · Ollama · spaCy | Mar 2026
An AI security pipeline that intercepts and sanitises LLM prompts before inference — blocking jailbreaks, prompt injections, and PII leakage at the source.
| Architecture | 3-stage pipeline: injection detection → PII analysis → policy engine |
| PII Coverage | Custom PatternRecognizer + context-aware boosters (+0.18) across 7 entity types |
| Injection Defence | Dual-tier regex detector — 14 patterns, 1.12× compound multiplier |
| Threshold | Optimised F1 score at 0.55 threshold |
| Inference | Integrated TinyLlama 1.1B via Ollama; negligible pipeline overhead confirmed |
🏥 Wasfa Diagnostic LMS | Flutter · Supabase · GoRouter · Flutter Quill | 2025–2026
Cross-platform LMS (web, mobile, desktop) for a live diagnostic centre — currently serving real patients in production.
| Scale | 400+ patient intakes/day digitised |
| Speed Gain | ~65% reduction in registration time via MRN/name/phone search + auto-fill |
| Report Editor | Flutter Quill + PDF export; clinician turnaround cut to <5 minutes |
| Auth | GoRouter + Supabase — 99%+ uptime, <2 sec median login |
| UI | Material 3, fully responsive across web, mobile, and desktop |
🗜️ Compresso | C++ · Qt Widgets · CMake · Multithreading | May 2025
Folder-based file compression tool implementing Canonical Huffman Coding — from scratch, no libraries.
| Compression | 45–60% file size reduction via optimised character frequency analysis |
| Engine | Multithreaded processing of 100+ files with custom binary format |
| Integrity | Header structures guarantee decompression accuracy |
📡 Eagle Eye – Proximity Radar | C · Arduino · Processing IDE | Dec 2024
Hardware-software radar prototype bridging embedded systems with a real-time GUI.
| Sensor | HC-SR04 ultrasonic + 180° servo motor |
| Interface | Real-time radar GUI via Processing IDE over serial at 9600 baud |
| Alert System | 4-band proximity zones (LEDs + buzzers at <10, 10–20, 20–40, >40 cm) |
🎹 ePiano | C++ · SFML | Jun 2024
Real-time virtual piano simulator with audio synthesis across 4 octaves and 48 keys.
| Latency | <10 ms audio synthesis latency |
| Keys | 4 octaves, 48 keys, event-driven architecture |
| Reception | 200+ downloads · 4.2/5 rating on university forum |
🗣️ Web Orator | JavaScript · Web Speech API | Aug 2024
Accessibility-focused TTS web app — built as a community service project for individuals with speech impairments.
| Impact | Deployed at a special needs centre; served 50+ users |
| Feedback | 90% positive rating |
| Coverage | 10+ languages, adjustable speech rate |
| Project | Stack | Highlights |
|---|---|---|
| 🏋️ Flutter Expense Tracker | Flutter · SQLite · FL Chart | 1000+ transaction records, visual budget analytics |
| 🍽️ Flutter Meals App | Flutter · Dart | Recipe categorisation, favourites, advanced filter system |
| 🌤️ Weather Dashboard | JavaScript · OpenWeatherMap API | Live data, condition-based theming, error handling |
| 🧩 Maze Puzzle Game | C++ · SFML · DFS | 15+ difficulty-scaled levels, DFS pathfinding, local leaderboard |
| 🏥 Medical Info System | C++ | Patient record management, structured data storage |
| 🥇 Award | 📋 Details |
|---|---|
| 🏆 ICPC Asia Topi Regional 2025 | Team Lead · The Pull Pirates · Ranked 73rd / 200+ teams |
| ⚡ NESCON 2025 Speed Programming | National-level · Team The Pull Pirates |
| 🎖️ Rector's Honours List | Nominated for academic excellence — Fall 2025 |
| 🏅 Merit Scholarship | Consecutive merit-based scholarships — CGPA 3.9 · 2024–Present |
|
Bahria University Islamabad | Feb 2026 – Present Selected via departmental nomination to represent Bahria University in global initiatives and facilitate student engagement in international scholarships and exchange programmes. |
Team: The Pull Pirates | 2024 – Present Founded the team from scratch. Conduct weekly algorithmic training sessions. Serve dual role as coach and active contestant at ICPC Asia Regional and national-level competitions. |
| 🎓 Certificate | 🏢 Issuer | 🔗 |
|---|---|---|
| Flutter and Dart: Developing iOS, Android & Mobile Apps | IBM via Coursera | Verify ↗ |
| Mobile App Notifications, Databases & Publishing | SkillUp via Coursera | Verify ↗ |
| Getting Started with Flutter Development | Google Cloud via Coursera | Verify ↗ |
Building things that matter — one commit at a time. 🚀