Hi !

• 👋 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), machine learning (Scikit-learn, PyTorch) with Python and SQL and API development (FastAPI) ;
  • familiar with DevOps/MLOps (Docker, CI/CD with GitHub Actions, GitLab CI, unit testing with pytest), GCP cloud (Big Query, Cloud Run, Vertex AI), Gen AI (Langchain, Haystack), vector databases (Chroma, Weaviate, PgVector), Front-end development (Vue.js, React.js) and agile development methods (Scrum, Kanban) ;
  • familiar with Infrastructure as Code with Terraform.
• 📫 Reach me by mail or Linkedin

Below are some projects I've worked on.

SightCall Q&A

Stack : Python (FastAPI, pytest), TypeScript (React.js), Haystack, PostgresSQL, Docker, Terraform, Github Actions

A web application for interacting with a RAG-based chatbot that answers questions about SightCall from their website.

• API
• Web app

Ask NERON

Stack : Python (FastAPI, pytest), TypeScript (Vue.js), Chroma DB, Docker

A chatbot web application which can answer question about eMush with Retrieval-Augmented Generation (RAG) from curated documents.

• Ask NERON API
• Ask NERON web app

cmnemoi-learn

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

eMush

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 :

• 2000+ users (150+ 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

OSS contributions

• LiveSplit autosplitters (https://github.com/cmnemoi/NuclearBlazeAutoSplitter)
• refactoring.guru
• srcomapi
• visions
• V tensor library

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

Predicting NYC Taxi Fare

Stack : Python (Pandas, Seaborn, Streamlit, scikit-learn, pytest), GCP, GitHub Actions

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)