🌟 About Me:

Currently enrolled as a Computer Science student at Warsaw University of Technology & Deep Learning Engineer.

🚀 I'm currently working on:

Generating and identifying deepfake alterations in video content

• In my master's degree project, I am tackling the challenges and opportunities presented by deepfake technology. The core of my research involves developing sophisticated methods to both create and detect deepfake videos, aiming to preserve the authenticity of digital media. I am leveraging 'TensorFlow' and 'PyTorch' as the primary machine learning frameworks, and 'OpenCV' for handling video data. A key focus is on utilizing Generative Adversarial Networks (GANs) to realistically generate and effectively identify video forgeries. Additionally, Transformer models are explored for their potential to enhance the generation and detection processes due to their advanced pattern recognition capabilities. The effectiveness of these techniques is rigorously tested using comprehensive datasets such as FaceForensics++.

Enhancing Continual Learning Through Classifier Calibration with Regularization

• This project focuses on employing various classifier calibration methods to enhance continual learning models, specifically using the Learning without Forgetting (LwF) technique. Classifier calibration is crucial for ensuring that a model's predicted probabilities accurately reflect the actual likelihood of events, a key factor in maintaining prediction reliability. Recent studies have shown that effective calibration can significantly enhance knowledge distillation processes. Based on these findings, we are applying these methods to investigate their potential to boost the performance of continual learning systems. We are currently using the FACIL GitHub repository as our primary resource for implementing the LwF method, leveraging both innovative and tried-and-true approaches to evaluate how they improve learning outcomes.

Diffusion model aimed at generating mathematical functions

• This project explores the innovative use of diffusion models to generate diverse mathematical functions, expanding the potential of generative AI in mathematical research. By adapting diffusion techniques, traditionally used for image generation, we aim to create models that iteratively refine randomness into structured mathematical expressions. The focus is on developing models that can propose novel functions, potentially offering new solutions or insights into complex mathematical problems. These functions are analyzed for their validity and utility, integrating advanced machine learning frameworks like TensorFlow or PyTorch, alongside mathematical tools such as MATLAB or SciPy for rigorous evaluation.

✨ I'm looking to collaborate on:

• Projects in Deep Learning and Data Science: I am keen to engage in projects that challenge and expand my current understanding and capabilities. If you're working on innovative projects or need an enthusiastic beginner's perspective in your team, I'm ready to dive in and contribute where I can, while learning from more experienced individuals.

🤝 I'm looking for help with:

• Techniques in Model Efficiency and Robustness: As a beginner, my goal is to understand the nuts and bolts of making deep learning models more efficient and robust. I’m looking for opportunities to assist on projects or research that focus on these areas. If you have the patience to guide and mentor, I’m eager to learn and help carry the workload.

🌟 Why Collaborate with Me:

• Mutual Learning: While I bring a fresh perspective and unbounded enthusiasm, collaborating with me is an opportunity to exchange knowledge in a way that enhances our mutual understanding of deep learning challenges and solutions.
• Grow Together: Join me in a collaborative journey where we both grow—me learning from the ground up, and you by refining your mentoring and leadership skills through guiding a newcomer.
• Expand Our Network: Working together opens doors to new connections and broadens our professional network, potentially leading to more collaborative opportunities and community engagement.

🌱 I'm currently learning:

• Advanced concepts in diffusion models and their applications in generative tasks, exploring how these models can innovate in areas like art generation and scientific simulation.
• State-of-the-art methods in neural networks, focusing on the latest advancements in architectures such as Transformers and their applications in natural language processing and beyond.
• Techniques in big data analytics, particularly how to efficiently process and extract value from large datasets using modern data science tools and machine learning algorithms.
• Ethical AI practices, ensuring that the models I develop are fair, transparent, and beneficial to society.

🔍 Ask me about:

• My experience developing a deep learning model for keyword detection in spoken sentences, including the challenges I faced, the technologies I used, and the successes we achieved in enhancing speech recognition capabilities.
• My work on solar panel detection, exploring how we leveraged machine learning to accurately identify and analyze solar panels from aerial imagery.

🔒 Note on Private Repositories:

Some of my projects are currently set to private either here on GitHub or on my university's private GitLab due to ongoing work and the sensitive or unfinished nature of the content. I am dedicated to making these projects available to the community once they are fully developed and ready for public sharing. Thank you for your understanding and patience!

⚡ Fun fact:

I not only enjoy delving into the intricacies of technology but also relish a good match of tennis to rejuvenate my mind and body

💬 Reach Out:

If you’re interested in mentoring a budding deep learning engineer or need an extra hand on your project, let’s connect! Message me here on GitHub or find me on LinkedIn to discuss potential collaborations.