Freelancer profile translated to English.

Back to original language

Description

Do you have data but unreliable, difficult-to-explain, or hard-to-deploy models?

I act as asenior data scientistto transform your data intorobust, actionable, and decision-oriented predictive models.

I assist companies, startups, and data teams with:

*Advanced data science & statistics

*Machine learning(classification, regression, clustering)

*Predictive modeling and time series(forecasting, scoring)

*Model validation, robustness, and interpretability

*Structuring, cleaning, and exploiting complex data

I primarily work inPython(pandas, NumPy, scikit-learn, statsmodels) and deliverclear, documented, and directly usable results.

**Typical deliverables**:

Ready-to-use predictive models
Clean and documented Python notebooks and scripts
Interpretable statistical analyses for decision-makers
Actionable data-driven recommendations

My differentiator: ahighly rigorous approach rooted in statistics**, combined with **pragmatism focused on business impact.

No unusable "black box" models: priority given to reliability, understanding, and decision-making.

Industry field of expertise

Languages

French
Native or bilingual
English
Fluent
Arabic
Native or bilingual

Workplace preferences

Can work on-site

Paris (up to 50km)

Capua
Data Scientist / Computer Vision Engineer
June 2025 - December 2025 (6 months)
Paris, France
Development of a real-time computer vision model for analyzing weightlifting movements (squat, deadlift) within a mobile application.

Mission focused on the design, optimization, and industrialization of machine learning models compatible with strong mobile performance constraints.

Design of computer vision models for posture and movement analysis
Consideration of biomechanical issues related to weightlifting exercises
Training and optimization of lightweight and high-performance models
Conversion and deployment of models to TFLite and ONNX
Seamless integration into a Kotlin Multiplatform architecture
Delivery of clean, documented, and production-ready Python code
Result: Fast, robust models directly usable in real-time on mobile, aligned with the application's product and technical constraints.
Data Science Machine Learning Computer Vision Time Series Analysis Python
RISQ/ALM, Société Générale,
Data Scientist / Statistical Modeling
July 2020 - June 2025 (4 years and 11 months)
Paris, France
Design, validation, and improvement of statistical and predictive models applied to customer and financial data.

Granular and aggregated data analysis
Statistical and behavioral modeling
Customer segmentation (clustering)
Robustness tests, validation, and sensitivity analysis
Development of Python libraries to automate analyses and calibrations

Production of clear, usable, and documented deliverables, presented to technical and non-technical stakeholders.
Python Critical Analysis & Synthesis Skills Machine Learning Statistics Applied Mathematics
Hellebore Capital,
Data Scientist / Statistical Research
April 2019 - September 2019 (5 months)
London, UK
Advanced data science project involving event and time series modeling.

In-depth statistical analysis of time series
Modeling using stochastic processes (multivariate Hawkes)
Development of a comprehensive Python package (analysis, calibration, visualization)
Numerical optimization and model validation

Strongdata science R&Dcomponent, full autonomy, and structured reporting of results.
Critical Analysis & Synthesis Skills Machine Learning Time Series Analysis and Forecasting Data Visualization Statistical Modeling