Building production-ready cross-platform applications and intelligent systems in Audio AI, Speech Processing, and Computer Vision.

• 📱 Flutter Developer building production-ready mobile, tablet, and desktop applications.
• 🎙️ Audio AI Engineer working on speech processing, phoneme recognition, alignment, segmentation, and duration estimation.
• 👁️ Computer Vision practitioner with experience in Arabic handwritten manuscript recognition and image-to-text systems.
• 🧠 Interested in deep learning, multimodal AI, OCR, ASR, and applied AI systems.
• 🏗️ Experienced with Clean Architecture, BLoC, REST APIs, CI/CD, PyTorch, TensorFlow, and production-oriented development.

Audio AI research project for phoneme recognition, phoneme boundary detection, duration estimation, and selected Tajweed-related analysis using Wav2Vec2, Conformer, CTC, and custom Quran-aware phoneme representations.

View Repository

Image-to-text system for recognizing Arabic handwritten manuscripts using image feature extraction and a Transformer decoder.

The project explores traditional computer vision preprocessing and deep-learning-based visual recognition, including CNN feature extraction, MobileNet, EfficientNet, and attention-based decoding.

View Repository

Open-source Flutter e-commerce application featuring authentication, product browsing, cart, wishlist, orders, checkout, reviews, local caching, and Appwrite backend integration.

View Repository

Educational implementation of neural networks directly in Dart, including:

• Multilayer Perceptron
• Elman recurrent network
• Manual forward propagation
• Backpropagation
• Weight updates
• Dataset preprocessing
• Model persistence

View Repository

Sudoku generator and logical solver implemented in Prolog with candidate elimination, uniqueness checks, pair strategies, and puzzle generation.

View Repository

I'm currently focused on:

• building reliable Flutter applications
• developing speech and audio AI systems
• working on phoneme-level speech analysis
• exploring computer vision and OCR
• improving deep-learning model design and evaluation
• building technically strong public AI projects

For the strongest profile presentation, I recommend this order:

1. quran-recitation-analysis
2. arabic-manuscript-ocr
3. flutter-ecommerce-shop
4. neural-networks-from-scratch-dart
5. prolog-sudoku-generator-solver

The sixth slot can stay open for your next strong project.

• 🌐 Portfolio: mahmoudbannan.com
• 💼 LinkedIn: linkedin.com/in/mahmoudbannan
• 💻 GitHub: github.com/Mahmoudmnb