I'm a Data Scientist & ML Engineering specialist at Decathlon France on a one-year professional training contract, passionate about transforming data into actionable insights and building production-ready ML solutions. Currently pursuing my Master 2 in Data Science at JUNIA, I specialize in developing end-to-end machine learning pipelines that drive business value.
Following a successful 3-month internship at Decathlon Belgium where I developed an automated sales forecasting pipeline, I'm now continuing my journey with Decathlon France to deepen my expertise in DS/ML engineering and production systems.
- Design robust and scalable DS/ML solutions - Creating architectures that scale with business needs
- Develop, validate, and implement DS/ML solutions - End-to-end model development from POC to production
- Monitor, maintain, and continuously improve DS/ML solutions - Ensuring optimal performance and business value
- Understand and define business problems and opportunities - Translating business challenges into data science solutions
- Communicate insights, document solutions, and foster data literacy - Bridging the gap between technical and business teams
- World Junior Freestyle Football Champion 2019
- Vice-Champion of France 2024
- Top 6 Europe 2024
- Founder of Nolan-Free Association - Promoting sports and creativity
During my 3-month internship at Decathlon Belgium (2025), I designed and implemented an end-to-end automated forecasting pipeline, replacing manual processes with a scalable ML solution that predicts sales across 60+ sports categories.
- Improved Accuracy: Improved forecast accuracy by ~10-15% compared to previous manual methods
- Automation: Reduced forecasting time from several days of manual work to ~10 minutes of automated processing
- Scalability: Successfully handles 60+ sports categories with parallel processing using joblib
- Model Performance: Prophet outperformed other models (LightGBM, XGBoost, Chronos) with 66.7% win rate across evaluation metrics
- Business Impact: Replaced manual Google Sheets processes with automated, reliable weekly forecasts
| Component | Technologies |
|---|---|
| ML Models | Prophet, LightGBM, XGBoost, Chronos-Bolt |
| Data Processing | PySpark, Pandas, NumPy |
| MLOps | MLflow, Databricks, Airflow |
| Cloud Infrastructure | AWS (S3), GCP (Vertex AI) |
| Feature Engineering | Weather data integration, Holiday calendars, Advanced time series features |
| Deployment | Docker, CI/CD pipelines, Automated reporting |
| Category | Technologies |
|---|---|
| Data Science & Analytics |         |
| Machine Learning & MLOps |         |
| Cloud & Big Data |            |
| Programming Languages |         |
| Web Development |         |
| Databases |         |
| Tools & DevOps |     |
I'm always open to discussing:
ML/AI Projects | Innovative Ideas | Collaborations | Data Science Opportunities
"Turning data into decisions, one model at a time"
Made with passion by Nolan Cacheux | © 2025