Mostafa Aly

Specific Areas of Interest: I am particularly interested in computer vision and natural language processing (NLP). My current work focuses on integrating these fields to create models that can translate textual descriptions into 3D representations, demonstrating the power of generative AI models for multi-modal tasks. Additionally, I have experience working with attention models in NLP and deep learning for image generation. Technologies/Frameworks: I primarily work with PyTorch for model development and TensorFlow 1.x for specific models like PRNet. I also utilize Hugging Face Transformers for fine-tuning Stable Diffusion models, and Gradio for building interactive model interfaces. In deployment, I use FastAPI and Hugging Face Spaces for hosting models. I'm also proficient in MLOps tools like MLflow and Hugging Face Hub, which enable streamlined model management and versioning. Applications: My work in generative AI has led to applications in image generation and 3D modeling, particularly in transforming text prompts into 3D facial models. I am fascinated by applications in creative AI (such as art generation), computer-aided design, and virtual reality, where the combination of text-to-image and image-to-3D technologies can unlock new possibilities. Research Interests: I am actively exploring how generative AI models, such as diffusion models and Generative Adversarial Networks (GANs), can improve the realism and diversity of generated content. My research interests include: Improving the efficiency of multi-modal models that combine text, image, and 3D data. Exploring methods to further integrate generative AI into professional workflows, particularly in areas like 3D asset generation for gaming, simulations, and digital art. Investigating how generative AI can be used ethically, focusing on issues like data privacy and fairness in AI-generated content. In addition to the project work, I’ve completed the AI & Data Science Track, where I studied key topics such as prompt engineering, NLP with attention models, GANs, and Generative AI models. I also gained hands-on experience with MLOps tools, particularly MLflow and Hugging Face.