Freelancer profile translated to English.
Back to original languageAbout Frédéric
👨🏻💻 My profile:
Results-oriented Data Scientist, I support companies from understanding business needs to deploying Machine Learning models, transforming data into high-impact business decisions.
Specialized in Machine Learning applied to time series, with a strong MLOps orientation, I cover the entire chain: scoping, modeling, industrialization, production deployment, CI/CD, monitoring, and documentation.
With over 3 years of experience in Data Science and 10 years of Python practice, I have led 4 successful projects, delivering robust and maintainable solutions.
🔧 Technical Expertise
🧪 Data Science & Machine Learning
- Supervised Machine Learning: regression, classification (churn, scoring), anomaly detection
- Time Series & Forecasting: Gradient Boosting (XGBoost, LightGBM, CatBoost), SARIMAX, Prophet, LSTM, GRU
- Clustering: K-Means, DTW
- Feature engineering, model evaluation (ROC-AUC, precision/recall, MAE, RMSE), interpretability
🏗️ MLOps & Industrialization
- ML Pipelines: feature preparation, training, deployment
- Production Deployment: Python APIs (FastAPI, Django)
- CI/CD, versioning, monitoring, reproducibility
☁️ Data Engineering & Cloud
- Data Pipelines: ingestion, transformation, orchestration (Pandas, Polars, Airflow, dbt)
- AWS (S3, compute, managed services)
- SQL / NoSQL Databases
- Large-scale data processing (Spark)
📊 Data Apps & Visualization
- Streamlit, Dash
- Exposure via REST APIs
⚡️ Energy & Industry
- Modeling and optimization of energy systems
- Consumption analysis (residential, commercial, industrial)
- Oil & Gas, low-carbon energies
- Operations Research
🤝 Soft Skills
- Explaining technical issues clearly
- Autonomy, rigor, delivery in complex contexts
- Leadership, excellent product / IT / business relationship
- 360° vision (research, startups, large corporations)
French
Native or bilingual
English
Fluent
German
Fluent
Can work on-site
Paris (up to 50km)
Experience
- TotalEnergies TOTSAData Scientist - Software EngineerENERGY AND UTILITIESNovember 2024 - December 2025 (1 year and 1 month)Paris, France🎯Mission:
- Development of an automated tool for analyzing non-routine flaring causes across all assets, combining time series processing in SQL and Python with anomaly detection techniques. Design of a Streamlit interface to visualize flaring trends, compliance with flaring policy, and statistical indicators (e.g., z-score) to identify assets requiring operational adjustments.
- Design and development of a tool for estimating GHG emissions related to venting on the Barnett asset, based on end-to-end processing of large time series datasets and rule-based integration with the BEST system.
- Product Owner role for LCS Advisor, the group's low-carbon solutions catalog, including defining functional specifications and overseeing the implementation of an internal RAG-based chatbot for internal decision support.
- FLEXUSData Scientist - Machine Learning - Time Series ForecastingENERGY AND UTILITIESJune 2024 - October 2025 (1 year and 4 months)Paris, France🎯Mission:
- Development of end-to-end time series forecasting pipelines to support Flexus's role as a flexibility aggregator for RTE. Design of a data workflow based on dbt and structuring of all customer consumption data in an SQL data warehouse hosted on AWS, enabling reliable 1 to 7-day forecasts.
- Customer segmentation using K-Means to group consumption profiles, extraction of seasonal patterns and exogenous factors using GAMs, and training of XGBoost models for multi-horizon forecasts on heterogeneous customer typologies, with an error below 10% after full hyperparameter optimization.
- Deployment of the forecasting pipeline on AWS with CI/CD practices, infrastructure automation, and production-level monitoring, ensuring the solution's scalability, maintainability, and robustness.
- FlunovaData Scientist - Software EngineerENERGY AND UTILITIESMay 2023 - May 2024 (1 year)Paris, France🎯Mission:
- Design and deployment of a full-stack web application (Django, JavaScript, SQL) integrating a 30-question building questionnaire, with the implementation of clear business rules and a modular software architecture inspired by Domain-Driven Design (DDD) and hexagonal architecture principles. Development of an energy performance estimation engine based on XGBoost, trained on public national datasets, and integrated into the platform to support user acquisition and conversion.
- Contribution to business strategy through user needs analysis, exploration of customer acquisition channels, and establishment of partnerships within the energy renovation ecosystem, positioning Flunova as a delegated operator. Deployment of the complete application and its prediction pipeline on a VPS.
Reviews
Recommendations
Be the first to recommend Frédéric
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Engineering DegreeCentrale LyonEcole d'ingénieur généraliste - spécialisation dans la data
- Double Degree - MasterTechnical University of MunichData science, Optimisation, modélisation de système d'énergie