AI student at Jordan University of Science and Technology (JUST), focused on deep learning, NLP, and multimodal machine learning.
My graduation project: a multimodal dual-branch deep learning architecture for MENA stock market forecasting. It fuses bilingual (Arabic/English) geopolitical and financial news with numerical time-series and technical indicator data, targeting stock exchanges in Jordan, Saudi Arabia, Egypt, Kuwait, and Qatar.
- Bilingual news + market data fusion using AraBERT and a Stockformer-based encoder
- Event-driven filtering (MIL) to isolate market-moving news from noise
- Full data pipeline covering ~300 companies across MENA exchanges
Python · PyTorch / TensorFlow · Keras · AraBERT · NLP · Flask · Docker · Pandas / NumPy · scikit-learn
- emotion-classification-roberta — Emotion classification using a fine-tuned RoBERTa model
- unet-image-enhancement — Image enhancement using a U-Net architecture
- RNN- — Recurrent Neural Network implemented from scratch using only NumPy
- mobile-price-classifier-mvc — Neural network predicting phone price range from specs, built as a clean Flask MVC app with Docker support
Building AI agents with the smolagents framework on Hugging Face Spaces.