Freelancer profile translated to English.
Back to original languageAbout Hermann
Do you want to leverage your data to improve your forecasts, automate your analyses, detect anomalies, or optimize your decisions?
I will help you design Data Science and artificial intelligence solutions tailored to your business needs.
I specialize in:
- Developing Machine Learning models;
- Forecasting sales, costs, risks, or performance;
- Detecting anomalies and weak signals;
- Optimizing industrial and operational processes;
- Automating data analyses and processing;
- Evaluating and improving existing models.
My approach combines business understanding, statistical analysis, and Python development. Before building a model, I clarify the need, assess data quality, and define metrics to measure the value created.
I have worked on projects in the energy, industry, mobility, and automotive sectors, particularly focusing on forecasting, optimization, maintenance, anomaly detection, and risk analysis.
I can assist with:
- Data audit or diagnosis;
- AI feasibility study;
- Prototype or proof of concept;
- Predictive model;
- Data preparation pipeline;
- Python automation;
- Validation, documentation, or industrialization of a solution.
I am co-founder of **Quasarzero**, a studio specializing in AI and Data Science. I personally handle each mission with a clear objective: to deliver a reliable, documented, and usable solution for your teams.
English
Fluent
French
Native or bilingual
Remote only
Primarily works remotely
Experience
- BMW franceData Scientist ConsultantAUTOMOBILEOctober 2025 - June 2026 (8 months)Paris, FranceForecasting maintenance costs & modeling residual valueLeading predictive modeling for maintenance costs: building regression and probabilistic models to forecast long-term costs and calibrate warranty strategy thresholds, with the goal of optimizing reliability.Continuing and industrializing the work initiated in 2024 on residual value slope analysis and vehicle pricing models (Premium & BEV segments).Technical Stack
- Modeling: Python (scikit-learn, statsmodels) — regression, probabilistic models, threshold calibration
- Data Engineering: SQL, data preparation and aggregation pipelines for vehicle/maintenance data
- MLOps: Docker, CI/CD for pipeline deployment and reproducibility
- AWS Cloud: ECR for image registry, App Runner/Fargate for scoring service deployment, S3 for data and artifact storage
Key Results- Maintenance cost models used to adjust warranty strategy
- Methodological continuity on residual value analysis (Premium & BEV), with a transition to an industrialized and reproducible pipeline
- La Française des Jeux (FDJ)Data Scientist Consultant — Optimizing player retentionSeptember 2024 - September 2024Paris, FranceDesigning a predictive system to identify players at risk of disengagement early on, enabling targeted retention actions on the most strategic profiles and improving the effectiveness of digital campaigns.
- Churn models (Python + SQL) deployed in production to prioritize retention actions on digital channels
- Player profile segmentation (persistent vs. ephemeral) based on their gameplay, to refine targeting
- Tracking framework over time to ensure models remain reliable month after month, preventing silent performance degradation in production
- Adjustment of decision thresholds to balance targeting accuracy and action volume, with direct production deployment on pipelines
- Prospective study on dynamic allocation algorithms (bandits) to optimize bonus distribution
- BMW France,Data Scientist ConsultantJanuary 2024 - July 2024 (6 months)Paris, FranceRevamping the vehicle pricing model for Premium & BEV segments, with a 12% increase in accuracy compared to the 2023 baseline — directly utilized by business teams for their pricing decisions.
- Vehicle depreciation analysis (SAS + Python) to refine pricing grids for Premium & BEV segments
- Rigorous comparison of several modeling approaches (linear regression, Ridge, polynomial, splines, Linear Forest) to select the most performant one
- Delivery of pricing grids by mileage × age, integrated directly into the sales teams' decision-making tools
Recommendations
Be the first to recommend Hermann
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
- Double MSc in Statistical Engineering and Data Science & in StatisticsSorbonne University – ISUP (Paris Institute of Statistics)2020
- Financial Risk Management (FRM) CandidateGARP2026Financial Risk Management (FRM) Candidate