Oussama Sounti

Data scientist specialised in applied machine learning and predictive modelling, with hands-on end-to-end experience both in industry and freelance from geospatial feature engineering and satellite remote sensing to building and rigorously validating real-estate valuation models. Strong foundation in business intelligence and software development, with particular attention to result reliability and data-leakage prevention.

Data Scientist associate at Yakeey (2026-02 – 2026-07)

• Building-age detection from satellite imagery: designed a multi-sensor fusion method (Google Earth Engine) to estimate construction year, validated against reference data.
• Points-of-interest (POI) scoring: designed a geospatial pipeline scoring density, diversity, and proximity to essential amenities (shops, transport) around each property, used as input features for the valuation model.
• Automated valuation model (AVM): designed and compared several machine learning models (RandomForest, ExtraTrees, XGBoost, LightGBM, CatBoost) for automated real-estate price estimation, achieving a significant error reduction over the existing baseline method.
• Implemented a strict time-based validation protocol (training on past data, testing on more recent data), with rigorous data-leakage control and automated hyperparameter tuning (Optuna).
• Added explainability (SHAP) and natural-language justifications for each price estimate.
• Estimates, along with an explanatory report, are shared by the company with partner banks as part of their property valuation process.

Data Analyst at Atass Recycler (2024-04 – 2024-12)

• Exploratory data analysis and development of machine learning models.
• Created visualisations and formulated recommendations; integrated models into existing systems.

Software Developer at Groupe AL OMRAN Beni Mellal Khenifra (2023-04 – 2023-11)

Web application for managing promotional email campaigns

• Managed client records and payment information (due dates, amounts).
• Targeted and sent email campaigns based on client behaviour; tracked their effectiveness.

IT Support Intern at Hôpital Ibn Zohr Marrakech (2022-06 – 2022-09)

Internship (PFA)

• Diagnosed technical issues, installed network and hardware equipment, and developed an archiving program.

Freelance Data Scientist at Self-employed (2024-01 – Present)

• Data science projects for Fiverr and direct clients, alongside the master's degree and the role at Yakeey: sentiment analysis, change detection, improvement of multi-criteria decision-making methods (AHP), and dimensionality reduction.

Freelance Developer at Self-employed (2022-01 – 2023-12)

• Web and application development for Fiverr and direct clients, alongside the DUT and professional Bachelor's degree (Laravel, ADO.NET, React, MongoDB, PostgreSQL administration).

Master's in Data Science for Smart Industry – École Nationale Supérieure d'Arts et Métiers, ENSAM Meknès, Morocco (2024-09 – 2026-06)

Professional Bachelor's in Infrastructure, Data Processing and Analysis (Big Data) – École Supérieure de Technologie, Fkih Ben Salah, Morocco (2023-09 – 2024-06)

University Diploma in Technology (DUT) in Computer Engineering – École Supérieure de Technologie, Béni Mellal, Morocco (2021-09 – 2023-06)

Baccalaureate in Physical Sciences and Chemistry – Lycée Demnate, Demnate, Morocco (2020-09 – 2021-06)