Exciting news! I've been hard at work making improvements to the workbench feature on this website and I'm thrilled to announce a significant upgrade: you can now export the underlying code for your data analysis and visualizations! This new functionality is a game-changer for anyone who wants more control over their data, allowing for better manual manipulation and deeper customization of both the data itself and its visual representation. This update provides a lot of flexibility. Get ready to take your data exploration to the next level by easily exporting the code and making it truly your own!

This month marks the start of my first semester at UCL, where I’ve officially begun my MSc in Data Science and Machine Learning. It feels exciting to dive into a program that combines both theory and practical skills, with a strong emphasis on applying methods to real-world problems. My modules this term include Introduction to Machine Learning, Introduction to Statistical Data Science, Blockchain Technologies, and Financial Engineering. Each of these areas opens up new perspectives: from understanding the fundamentals of ML models, to building statistical foundations for data-driven insights, to exploring the role of blockchain in decentralized systems, and applying quantitative methods to finance. I’m looking forward to the challenge, the projects, and the journey ahead as I build expertise at the intersection of data science, machine learning, and applied technology.

Over the past few weeks, I’ve been building a custom Machine Learning Workbench that helps me prepare datasets for Kaggle competitions and personal ML projects. The workbench now includes a CSV Analyser for quick stats and visualizations, a Categorical Encoder supporting label, one-hot, and target encoding (with editable mappings!), and a Feature Scaler that lets me choose scaling methods column-by-column. Each tab exports a cleaned CSV so I can iterate fast. The next steps will include missing value handling, correlation analysis, and a model runner tab to close the loop. The goal is to have an end-to-end, interactive UI for data exploration, transformation, and modeling — all inside a single app.

The Prompt Inserter Chrome extension has been officially published on the Chrome Web Store! Users can now quickly insert AI prompt templates into ChatGPT with the newly improved interface, clearer button layouts, and faster access to templates. The extension is fully compliant with all Chrome Web Store guidelines, including privacy justifications and permissions. Future updates will expand compatibility to AI services like Gemini and continue refining the user experience for seamless workflow integration.

After developing the Prompt Inserter Chrome extension to quickly insert AI prompt templates into ChatGPT, I focused on enhancing the user interface to make prompt selection and insertion even smoother. The extension now includes a more intuitive popup design, clearer button layouts, and faster access to user-defined templates. I am currently preparing the extension for publication on the Chrome Web Store ensuring all privacy justifications, permissions, and metadata meet their guidelines. Future updates will expand compatibility to AI services like Gemini and refine the UI further for seamless workflow integration.