Fancy seeing you here 😌 !

I am Maria Balos, a data scientist and user-centric designer based in Cambridge, UK. Nowadays, I work for VocalityAI, training TTS models, so you can find me most of the time behind a screen or next to a cup of coffee. In my free time, I enjoy solving problems on Leetcode or experimenting while working on personal projects. Please grab a coffee and feel welcome to this small corner of my work!

LeetCode & Github Stats

Right now I am involved in:

• Developing and deploying speech generative models for commercial applications, ensuring their efficiency, scalability, and seamless integration into real-world use cases.
• Improving my code skills with Leetcode.
• Implementing CNN architectures for image classification.

Last achievements:

• 2025/12/24: Finished and published the project: CON(e)VOLUTION-A Walkthrough From LeNet to Vision Transformers
• 2025/07/16: Presented Master's Dissertation for the Deep Learning and Generative AI Master's Degree at Datamecum called RAG-Driven Educational Assistant: Enhancing Learning through Contextual Q&A and Knowledge Testing. Please review the main repository by following the link provided.
• 2025/01/27: Started working at Vocality.AI
• 2024/12/01: Helping Santa Claus arrive in time for Christmas by solving problems in the AdventOfCode 2024. Check it out in Advent of Code
• 2024/11/01: I reached 200 solved Leetcode problems: Check it out in my Leetcode profile
• 2024/10/30: Ryanair time-capsule, reverse-engineering the Ryanair API to collect daily flight prices and train machine learning models to forecast price changes.
• 2024/10/17: Completed NLP HuggingFace course.
• 2024/05/28: Completed the first part of the "Practical Deep Learning" course by fast.ai
• 2024/05/17: Completed the "Advanced Learning Algorithm Course" by Andrew Ng in Coursera
• 2024/05/09: Winner of the DATAMECUM Datathon 3rd promotion competition. Winner of the Datamecum DATAMECUM Datathon 3rd promotion - Datathon competition. Capstone project won with an ensemble of the Random Forest and XGBoost predictions, please check out the [presentation video](https://youtu.be/kcrQabXlEoU?

Medium posts & Kaggle notebooks:

• CON(e)VOLUTION-A Walkthrough From LeNet to Vision Transformers
• Exploratory Data Analysis (EDA) for Python Programmers — Part 1
• Machine learning applied to the design industry: K-Means for image palette generation
• The Power of Decision Stumps

• Space Mission Analysis is a data exploration and data visualisation project where I applied most of the data visualisation libraries.

• Mohs Hardness Exploratory Data Analysis: Decision Stump (one layer decision tree) for a Kaggle competition, this placed me in position 598/1632 at the end of the competition. A decision stump presentation has been created to introduce Datamecum students to decision stumps.

• Datamecum Datathon - Capstone project competition between the third promotion students of the Intensive Program in Data Science by DATAMECUM, consisting of building a supervised model to predict a binary class. The exploratory data analysis consisted of:

  • checking for missing values.
  • handling duplicated values and ambiguous data.
  • exploring the relation between missing values and the target variable.
  • self-organising maps and correlation matrix used for correlation checks.

• Weever Watermark: Applying K-Means and arrays transformation to group the colours of a provided image and use the generated centroids to create a 10-colour palette into a GUI API created with Flask. Please have a look at the Weever Watermark DEMO or at the Machine learning applied to the design industry: K-Means for image palette generation article, where I explain the project.

• Datamecum Dataton - Capstone project competition between the third promotion students of the Intensive Program in Data Science by DATAMECUM, consisting of building a supervised model to predict a binary class.

• Typing Thunder: a speed-typing GUI app created to measure how fast the user type in one minute. Typing Thunder DEMO

• Morse Code: a command-line program where the dictionary loops, strings and functions in Python are applied. Morse Converter DEMO

• MochaMaps: a website that displays coffee shops and their facilities from a database by using: SQLite, SQLAlchemy, Jinja2, REST API, Flask API and Bootstrap-Flask. MockaMaps DEMO

• LinkedIn Toggler: Using Selenium for Python to automate recurrent LinkedIn tasks.

Thank you for visiting my GitHub! Feel free to have a deeper look at my repositories to find more specific projects. Please share any feedback, suggestions, or tips that you believe could help me grow and improve!

I am always happy for a coffee, a chit-chat or a discussion of any possible collaboration. Please drop me an email at mariabalos16@gmail.com or send me a message through my LinkedIn if you fancy any of those.