Welcome to my personal webspace! I'm Panos Kolyvakis, a Research Data Scientist with a PhD in materials science, particularly hyper-crosslinked polymers within the field of soft matter. With lecturing experience at the University of Edinburgh in Environmental Engineering and sustainability, I bring a wealth of knowledge in several scientific and engineering disciplines. An academically strong STEM individual with some business intelligence and data analytics experience and a profound passion for innovation and AI technologies. Dedicated to advancing predictive modeling techniques and committed to crafting high-quality code that addresses real-world problems. Continuously seeks to merge analytical rigor with creative problem-solving to drive forward the boundaries of what's possible.
Projects
AI to LaTeX PDF Generator
A web application leveraging Google Custom Search and OpenAI's GPT to create and format blog posts into LaTeX, compiled into downloadable PDFs. Streamlines professional document creation for all users.
Technologies Used: Python, Flask, Javascript, HTML, CSS, Google Programmable Search Engine, OpenAI's GPT, Web Scraping, Full-stack application, WebUI, LaTeX
SnippingTool-To-GPT models
A macOS Quick Action tool capturing screen selections, extracting text via OCR, and leveraging GPT models for insightful analyses. Features code editor-like functionality featuring code highlighting and tk GUI for GPT responses.
Technologies Used: Python, Tesseract OCR, OpenAI GPT, Bash Scripting, Automator for macOS
Spartan Engine Contribution
My engagement with the Spartan Engine showcases my dedication to open-source contributions and compiled language programming. I focused on augmenting the engine's physics capabilities, introducing floatation dynamics and movement physics. This effort enhances the realism and interactivity within the engine's rich graphical worlds.
Contributions: Physics for floatation, movement dynamics, open-source enhancement.
General Machine Learning
Exploring the breadth of Machine Learning through projects utilizing XGBoost, Decision Trees, and advanced techniques for Classification, Regression, and Time Series Forecasting. Includes hands-on applications of Neural Networks and Convolutional Neural Networks to solve real-world problems.
Technologies Used: XGBoost, Decision Trees, CNNs, Python, TensorFlow, Keras, Scikit-learn