Pierre Aumjaud
Data Analyst | Analytics Engineer | Data Scientist
Data-focused professional transitioning from an academic career after a four-year period of dedicated skills development. Proficient in data engineering, data analysis, and data science, with a proven ability to build and deploy machine learning models.
I specialise in:
-
Data Analytics
- Databases : PostgreSQL • DuckDB • MySQL
- Data visualisation : Tableau • Metabase • Streamlit • Plotly • Matplotlib • Seaborn
- Process monitoring : Grafana • Prometheus
- Data cleaning : Python • Numpy • Pandas
-
Data Integration
-
Data Science
- Predictive modeling : Python • Scikit-learn • Pytorch • Jupyter
- Anomaly detection
- Reinforcement learning : Stable-Baselines3
- Evolutionary optimisation
-
DevOps
- Versionning : Git
- Container : Docker
- CI/CD : Github Actions
- Unit test : Pytest
-
Interactive Dashboard with Metabase
Built an interactive dashboard with Metabase to analyse journalist fatalities in the world.
-
MLOps Pipeline Deployment
Deployed a machine learning model that predicts patient medical charges based on demographic and health data.
-
Performance Monitoring with Grafana
Monitored performance metrics of a predictive regression model using Grafana and automated alerts.
-
Deployment of a Large Language Model Web Application
Deployed a chatbot prototype to explore the capabilities of large language models. This project involved integrating the Llama 2 API from Replicate into a responsive front-end web application built with Streamlit.
-
Customer Data Cleaning with SQL
Processed raw customer data using SQL by removing duplicates, handling missing values, standardizing formats, and splitting columns for better analysis. Ensured data integrity and prepared it for actionable insights.
-
Data Visualisation with Tableau
A collection of data visualisation dashboards with Tableau.
-
Reinforcement Learning for Robotic Arm Control
Trained a reinforcement learning agent in a custom Gymnasium environment to solve a robotic reach task using PyBullet physics. This project demonstrates my ability to implement RL algorithms, simulate robotic systems, and optimize control policies for real-world applications.
-
Classifier Visualisation
Developed an intuitive Python tool to train, evaluate, and visualize decision boundaries of multiple classifiers (SVM, Random Forest, Logistic Regression) on 2D datasets. Implemented hyperparameter tuning to optimize model performance while providing visual explanations of model behavior and trade-offs.
-
Data Analysis and Regression Predictions with Python
Built a predictive model in Python to forecast residential home prices, applying machine learning with scikit-learn to solve a supervised regression problem.
-
Exploratory Data Analysis and Classification Predictions with Python
Developed a predictive model to identify factors influencing survival during the Titanic disaster. This project involved a complete machine learning workflow, from data cleaning and feature engineering to fitting a classifier.
-
Custom Reinforcement Learning Environments
Developed modular Gymnasium environments for training RL agents, integrating physics-based robotics simulations via PyBullet and ROS.
-
Multi-Objective Optimization in Python
A collection of practical examples and visualizations demonstrating evolutionary algorithms, constraint handling, and Pareto front analysis using the PYMOO framework.
