Arash keyhan

I've built end-to-end Python projects in machine learning, data analysis, and scientific computing, including California housing price prediction, MNIST digit classification, and projectile-motion simulation.

For my MSc thesis at Bu-Ali Sina University, I reconstructed the Universe's expansion history and linear growth of matter perturbations using Hubble parameter and growth-rate data. I implemented approximately 70% of the project in Python, applying train-test splits, polynomial regression, and cross-validation to minimize error against theoretical predictions.

My digit classifier, built with PyTorch on the MNIST dataset, achieved 92.03% test accuracy. I also developed a command-line bank management system using object-oriented programming and a housing-price pipeline using preprocessing, feature engineering, stratified sampling, regression, and model evaluation.

I'm currently studying Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, and I'm motivated to turn my analytical physics background into practical data and machine learning work. I learn independently, work hard, and aim to demonstrate my ability through results.