Hi, my name is
Abdullah Altaf.
I build intelligent systems.
I’m a software engineer specializing in building AI-powered systems, backend services, and data-driven applications. Currently, I’m focused on developing scalable Python-based solutions and Retrieval-Augmented Generation (RAG) pipelines.
About Me
Hello! I'm Abdullah Altaf, a developer who genuinely loves turning complex ideas into intelligent, living software. My journey in tech began with an insatiable curiosity for how systems work behind the scenes, which led me to pursue my BSCS at Information Technology University (ITU).
Today, my heart lies in applied Artificial Intelligence, backend engineering, and intelligent automation. Whether I'm building conversational RAG pipelines, training neural networks for traffic safety, or engineering custom VS Code extensions, I thrive on creating solutions that make a tangible difference. For me, software isn't just about writing code—it's about crafting seamless experiences that truly empower people.
From earning my Oracle AI certification and exploring advanced Model Context Protocols to tackling complex geospatial analytics, I'm constantly exploring new frontiers. I believe the best tech feels like magic, and I'm dedicated to building systems that bring that magic to life.
Here are a few technologies I’ve been working with recently:
- Fastapi
- Python & Shell Scripting
- Agentic Libs
- RAG & Vector DBs
- MySQL & Firebase
- Docker
Where I’ve Worked
AI/ML Intern (RAG Chatbot Developer)
@ Minhaj University
September - November 2025
- Developed a Retrieval-Augmented Generation (RAG) AI chatbot for the official university website to handle admission and general information queries in real time.
- Designed the end-to-end RAG pipeline, including document preprocessing, vector embeddings, and FAISS-based semantic retrieval.
- Integrated Large Language Models using the OpenAI SDK to generate contextual and accurate responses.
- Structured and optimized university data to improve retrieval quality and minimize hallucinations.
- Assisted in backend integration and deployment to enhance accessibility and automate student support.
Final Year Project
Capstone Project Spotlight
AI-Powered Traffic Violation Detection & Automated Enforcement System
We have developed an AI powered system which monitors the traffic with a single low-cost CCTV camera and automatically catches the rule violators like red light signal violation, wrong side driving, zebra-crossing violations, and riding without a helmet, etc. with number plate reading using an OCR model trained from scratch. The system then locates the hotspots for violations allowing the traffic police to know where to send wardens or put permanent cameras. Generating Car Traffic Is An Affordable Way To Make Roads Safer.
- YOLOv8
- OpenCV
- PaddleOCR
- StrongSORT
- Python
- PyTorch
- CSRT Tracker
- NumPy
- CCTV
- Kaggle
Some Things I’ve Built
Featured Project
Remote Adb - Vscode ExtensionA VS Code extension that provides a user interface to do development right in an Android device via ADB. This powerful tool brings your Android environment directly into your editor, streamlining debugging and testing workflows by bridging the gap between local development and remote device execution.
- TypeScript
- Shell
- Android
- Scripting
- VS Code
Featured Project
MCB Branch Placement Analysis
Conducted geospatial analysis to identify strategic locations for new MCB branches and ATMs based on competitor density, accessibility, and population distribution. Employed Kernel Density Estimation (KDE) and spatial joins to compute suitability scores and visualize results via interactive maps. Proposed top 3 data-driven locations for MCB expansion under both 'No Bank Area' and 'Rival Present' scenarios.
- Python
- Geospatial Analysis
- KDE
![MCB Branch Placement Analysis]()
Other Noteworthy Projects
view the archiveBlockchain Land Registry dApp
Designed and deployed a decentralized land registration platform enabling secure property transactions via Ethereum smart contracts.
Termux Over ADB
PowerShell automation toolkit that injects native Termux environment variables into interactive ADB sessions for zero-daemon toolchain execution.
Full-Stack Portfolio & Multimedia Hub
Full-stack portfolio platform featuring an interactive 60fps canvas vertical video feed and a rate-limited FastAPI confession backend.
Neural Network for Skin Tone Classification
Built a custom neural network from scratch to classify skin tones into categories Designed and trained a model using sigmoid and softmax activation functions with batch gradient descent.
PCA Face & Attribute Recognition System
Eigenface recognition engine built from scratch in NumPy, classifying faces, emotions, and facial attributes via eigenspace projection and reconstruction loss.
C++ Klondike Solitaire Engine
Object-oriented terminal Klondike Solitaire game featuring stack-based card manipulation, Fisher-Yates deck shuffling, and configurable draw modes.
Certifications
Oracle Cloud Infrastructure 2025 Certified AI Foundations Associate
Validation of foundational knowledge of Oracle Cloud Infrastructure (OCI) and its specific applications within the field of Artificial Intelligence.
Model Context Protocol: Advanced Topics
Advanced topics in Model Context Protocol for Agentic AI interactions.
Deloitte Australia - Technology Job Training
Completed a job simulation involving development and coding, and wrote a proposal for creating a dashboard.
Problem Solving (Intermediate) Certificate
Demonstrated intermediate problem-solving skills in algorithms and data structures.
What’s Next?
Get In Touch
Although I’m not currently looking for any new opportunities, my inbox is always open. Whether you have a question or just want to say hi, I’ll try my best to get back to you!
Say Hello