Software Engineering Student at NUST | AI & Machine Learning | Flutter | Full-Stack Web (MERN)
Second-year Software Engineering student at SEECS, NUST, Islamabad. I build across three domains: mobile (Flutter), full-stack web (MERN), and machine learning. Currently going deep on Flutter while maintaining applied ML work through real projects and Kaggle competitions.
- Location: Islamabad, Pakistan
- Portfolio: osamaayaz.vercel.app
- Email: usamaayaz475@gmail.com
- LinkedIn: Osama Ayaz
ML & DL: TensorFlow, Keras, PyTorch, Scikit-learn, OpenCV
Web Dev: React, Node.js, Express, MongoDB, TypeScript, Tailwind CSS
App Dev: Flutter, Java (Android)
Languages: Python, C++, Java, JavaScript, TypeScript, Dart, SQL
AI-powered telemedicine platform with real-time voice consultations using specialized medical agents,
symptom analysis, live transcription, and automated generation of structured medical reports.
Tech: Next.js, React, TypeScript, Tailwind CSS, LangGraph, LlamaIndex, Voice AI
Repo: github.com/usama-codes/mediassist
Full-stack platform for university communities with auth-only registration, multi-image item reporting,
ownership verification, and a gamified contributor leaderboard.
Tech: Next.js, TypeScript, React, MongoDB, NextAuth
Repo: github.com/usama-codes/lost-and-found-app
High-performance search engine indexing 1.39M+ Medium articles using an offline Python indexing pipeline and a low-latency C++ query engine.
- 521k-word lexicon with forward and inverted indexes split into 50 barrels for parallel query processing
- Tokenization and lemmatization via NLTK; ranking via TF-IDF
- Architecture: Flask API (ingestion) + C++ query engine (ranking) + React/TypeScript frontend
Tech: C++, Python, Flask, React, TypeScript, NLTK
Repo: github.com/usama-codes/search-engine
CNN-based malignant vs. benign classifier built independently during internship at Murabbi using the ISIC 2024 skin lesion dataset.
- Transfer learning on ResNet50 with custom color histograms, LBP texture descriptors, and contour-based shape features
- Addressed class imbalance via targeted data augmentation
- Results: 92% test accuracy · 89% recall · F1 score of 60%
Tech: Python, TensorFlow, Keras, OpenCV
Repo: github.com/usama-codes/ISIC-2024-Competition
Voice-activated AI assistant handling WhatsApp messaging, weather forecasts, news briefings,
system control (shutdown/restart/sleep/lock), application launching, and local music playback.
Tech: Python, Speech Recognition, Flask
Repo: github.com/usama-codes/Jarvis
Android app for managing sports ground reservations, waiting queues, and society event
tracking across the university.
Tech: Java, Android SDK
**Repo:"" github.com/OmgItsAmmer/festiva
ML/DL Intern — Murabbi (Feb 2024 – May 2024)
Worked independently on a cancer detection project applying transfer learning on ResNet50,
custom feature engineering, and CNN-based classification on real-world ISIC medical imaging data.
Flutter Intern - Cognitive Health Intl. (April 2026-Present)
Working on Flutter development under a highly skilled team.
B.E. Software Engineering — SEECS, NUST (2023 – Present)
FSc Pre-Engineering — Army Public School and College, Mardan (2021 – 2023)
- Flutter — primary focus, production-grade mobile development
- Kaggle competitions — applied ML practice
- Exploring Agentic AI