• 👋 I’m Charles-Meldhine Madi Mnemoi. I am a Data Scientist in Co-op by day and a full-stack developper for eMush by night.
• 🛠️ Skills
  • proficient in data analysis (Pandas, Matplotlib, Seaborn, Plotly) and machine learning (Scikit-learn, PyTorch) with Python and SQL ;
  • familiar with DevOps/MLOps (Docker, GitHub Actions, GitLab CI, unit testing with pytest), API development (FastAPI, Flask), GCP cloud (Big Query, Cloud Run, Vertex AI) and agile development methods (Scrum, Kanban) ;
  • acculturated to Gen AI with Langchain and vector databases (Weaviate) ;
• 📫 Reach me by mail or Linkedin

Below are some projects I've worked on.

cmnemoi-learn is a Python package which reimplements machine learning algorithms from scratch (using only numpy) with high quality development practices :

• unit testing with pytest
• code quality checking with black, pylint and mypy
• CI/CD pipeline with GitHub Actions to version and publish the package automatically to PyPI

Stack : PHP 8.3 (Symfony 6.4, PHPUnit, Codeception), Vue.js 3, PostgreSQL, GitLab, Docker, GitLab CI

eMush is an open source remake of Mush: the greatest space opera epic of Humanity, directly on your browser!

I am a full-stack developer for the project since July 2022.

KPIs :

• 1500+ users (100+ daily)
• contribution to 100 000+ lines of code

Missions :

• feature development, bugfixes and testing
• enhancement of CI pipelines
• implementing good practices (TDD, BDD, Clean Architecture)
• participation in discussions on project direction and features to be developed
• writing monthly news and patchnotes
• animating alpha tests

I've done the projects below when I was starting in Data Science and software engineering, they deserve a reboot now...

A web application to predict of the minimum duration of a video game session with Machine Learning.

Stack : Python (pandas, sklearn, matplotlib, seaborn, requests, beautiful soup, flask, streamlit)

• Developed a web application that estimates the minimum duration of a match within a 6-minute margin
• Extracted 6,000+ game sessions' data from APIs and web scraping with requests and BeautifulSoup
• Created new variables based on game session time, number of players, and the game's release year
• Built an API using Flask and the web application using Streamlit

Data Science project of Lille's Bachelor of Economics, which consists of participating in the Kaggle competition New York City Taxi Fare Prediction.

• Developed a web application that estimates the price of a ride within a $1.4 range
• Cleaned and analyzed a dataset with 340,000+ rows to remove outliers and noise from data with normalization
• Created new variables based on ride duration and destinations
• Built the web application using Streamlit
• Quality "CI" pipeline with git hooks and Github Actions (lint with Ruff, test with Pytest)

A web app allowing you to compare French Hip-Hop lyrics.

Stack : Python (Streamlit, Pandas, Plotly, Matplotlib, LyricsGenius, SQLAlchemy), MySQL, AWS RDS

• 21723 songs collected through LyricsGenius API
• MySQL database hosted on a AWS RDS DB instance
• Nice looking charts automatically generated from data and user entries
• Web app with Streamlit
• Quality "CI" pipeline with git hooks and Github Actions (lint with Ruff, test with Pytest)