Moeez.dev
AVAILABLE FOR OPPORTUNITIES · GUJRANWALA

Muhammad
Moeez

A Deep Learning & Full-Stack developer who builds end-to-end products — from NumPy-only neural nets to production-ready web apps.

Muhammad Moeez

01 · ABOUT

Building things that think and things that ship.

I'm Muhammad Moeez, a developer based in Gujranwala, Pakistan, currently pursuing a Bachelor of Information Technology at the University of Gujrat.

My work sits at the intersection of two disciplines: teaching machines to understand data, and building the interfaces people actually use to interact with it.

On the AI side, I've implemented neural networks from scratch with nothing but NumPy, and trained convolutional models like ResNet-18 for real-world medical imaging problems. On the engineering side, I build responsive front-ends with React, mobile apps with Flutter, and backends with Node and Flask.

Focus

Deep Learning, Computer Vision & Full-Stack Web/Mobile

Based in

Gujranwala, Punjab, Pakistan

02 · SKILLS

What I work with

Deep Learning & AI

  • → Building models for deep learning & statistical use cases
  • → Computer vision & NLP project experience
  • → Quantitative modelling for forecasting & time series
TensorFlowKerasPyTorchPythonNumPy

Full-Stack Development

  • → Responsive front-ends with React
  • → Mobile apps with Flutter, React Native & Kotlin
  • → Backends in Node, Express & Flask
HTML5CSS3JavaScriptReactNode.jsFlutter

Sharpening the edge

Solving problems consistently on competitive platforms

03 · PROJECTS

Selected work

A mix of deep learning experiments and full-stack products. Live activity from GitHub is further down the page.

MACHINE LEARNING

Face Recognition System

A from-scratch face recognition system built with a NumPy-only multilayer perceptron — no deep learning frameworks, just linear algebra and gradient descent, implemented by hand.

PythonNumPyJupyter
MEDICAL IMAGING

Alzheimer's Disease Detection

A deep learning pipeline that classifies Alzheimer's disease stages from brain MRI scans using a fine-tuned ResNet-18 convolutional network.

PythonResNet-18Computer Vision
FULL STACK

The Living Archive

A platform for documenting discoveries and urban exploration — built for explorers to log, tag, and share findings from abandoned places and hidden spots.

TypeScriptWeb App
MANAGEMENT SYSTEM

SpinNSpice

A restaurant management system for handling orders, menus, and day-to-day operations end to end.

AI · EDTECH

DirectionWise

A career-counselling tool that helps students explore paths and make informed decisions about their future.

FRONTEND

NeoShop

A demo e-commerce storefront showcasing product browsing, cart interactions, and a clean, modern shopping UI.

JavaScriptE-Commerce

04 · UPCOMING

What's next

Work in progress or on deck — updated as it happens.

IN PROGRESS

NUZM Tech — Full-Stack Website

A full-stack website build for NUZM Tech, covering the complete product from front-end to backend infrastructure.

Full StackWeb Development
IN PROGRESS

AI Humanizer

A tool that rewrites AI-generated text to read more naturally and human, for writers who want their AI-assisted drafts to sound authentically their own.

AINLP
WRITING

The Interplay between Mental Health and Student Life in Higher Education: An Examination of Contributing Factors and Resilience Strategies

A review paper examining how mental health and student life interact in higher education — the factors that put students at risk, and the resilience strategies that help them cope.

Review PaperMental HealthHigher Education
RESEARCH

Designing a Computer Game Using the Theory of Automata

A research paper exploring how finite automata and formal language theory can be used to design and drive the logic of a computer game.

Automata TheoryGame DevelopmentResearch Paper
RESEARCH

Alzheimer's Disease Stage Classification and Detection Using ResNet-18 with Grad-CAM: A Quantitatively Validated Explainability Approach

A research paper extending the Alzheimer's detection work with Grad-CAM-based explainability, quantitatively validating that the ResNet-18 model's attention aligns with clinically relevant regions.

ResNet-18Grad-CAMExplainable AIResearch Paper
IN PROGRESS

Jarvis — Mobile & Laptop Automated AI System

A voice-driven AI automation system for both mobile and laptop, aimed at handling everyday tasks hands-free — inspired by the classic sci-fi assistant.

AIAutomation

LIVE · GITHUB

Latest activity

    05 · CERTIFICATES

    Certifications

    Courses and certifications, completed and in progress.

    Certificates coming soon

    This section is ready to go — certificates will be added here shortly.

    06 · COMPETITIONS

    Competitions

    Competitions and hackathons applied to or entered.

    Nothing listed yet

    Competition entries will be added here shortly.

    07 · EDUCATION

    Academic path

    2023 — 2027

    University of Gujrat

    Bachelor of Information Technology

    Coursework spanning Explainable AI, Graph Machine Learning and Computer Vision. Also involved with the university's multimedia department, working on documentary films and interviews.

    2021 — 2023

    Punjab College

    Intermediate in Computer Science

    Core computer science foundations — Data Structures, Algorithms, DBMS, Operating Systems and Computer Architecture — alongside self-directed courses in Deep Learning, Data Science, Cloud Computing and Full-Stack Development.

    08 · ACHIEVEMENTS

    Milestones so far

    9+

    Public Repositories

    Shipped and open-sourced across machine learning, computer vision and full-stack web on GitHub.

    0 → 1

    Neural Net From Scratch

    Implemented a multilayer perceptron for face recognition using nothing but NumPy.

    3

    Live Deployments

    The Living Archive, NeoShop and DirectionWise — all built, deployed and publicly reachable.

    3

    Competitive Platforms

    Active problem solver on LeetCode, HackerRank and Kaggle.

    Multimedia

    Documentary Contributor

    Worked with University of Gujrat's multimedia department on documentary films and interviews.

    09 · CONTACT

    Let's build something worth shipping.

    Open to internships, collaborations and interesting problems in AI or full-stack engineering.

    OR SEND A MESSAGE DIRECTLY