Muhammad Osman

I am a Data Scientist with hands-on expertise in Machine Learning, Deep Learning, and AI model development. My proficiency lies in designing and deploying advanced algorithms, including classification, regression, CNNs, RNNs, Autoencoders, and Attention Models. I have a strong command of Python, TensorFlow, PyTorch, and Scikit-learn, which allows me to develop scalable, data-driven solutions that enhance business decision-making.

Throughout my academic journey, I have worked on various projects that showcase my skills in machine learning. For instance, I built a Transformer-based text summarization model during my Master's program at the University of Hertfordshire, where I implemented advanced techniques such as positional encoding and multi-head attention. Additionally, I developed a stock price prediction model using LSTM networks and a weather forecasting model using GRU networks, demonstrating my ability to apply deep learning techniques to real-world problems.

My experience also includes developing a Variational Autoencoder for face reconstruction and a CNN model for currency note classification. These projects not only highlight my technical skills but also my commitment to delivering interpretable results and practical applications in various domains, including financial security.