Data Scientist

Technical Skills: Python, SQL, R, Google Cloud, Tableau

Education

• MSc., Data Science

  University of Utrecht, the Netherlands (June 2023)

Work Experience

Data Engineer @ Societe Generale (Oct 2024 - Present)

• Implemented automated data quality controls using Python and Teradata, reducing manual work by 85%.
• Managed data warehouse changes through deployments that ensured timely and reliable data for reporting and analytics.
• Envisioned Oracle solution to efficiently parse 500M+ XML files, improving data accessibility for BI.

AI Automation Specialist @ Brandsom (Jan 2024 - May 2024)

• Used GPT‑4 Vision to automate information extraction from product images, saving ~60 hours of manual work monthly.
• Developed a tool using prompt engineering and Python to automate the process of generating descriptions for Amazon.nl and Bol.com based on products’ attributes, with a potential to reduce manual work by ~70 hours per month.

Integration Specialist @ Brandsom (May 2022 - Dec 2023)

• Used regression models to analyze price elasticity, providing clients with data‑driven insights into optimal pricing strategies.
• Led the development of an ETL (scraping‑airflow‑cloud‑dashboard), resulting in a new dashboard product for the largest client Asics.

Data Analyst @ Gajos (Jan 2021 - Mar 2022)

• Built interactive dashboards in Tableau to visualize the performance of the eshop and website for management review.
• Developed SQL stored procedures to optimize product import from Gajos’s suppliers, reducing monthly manual work by 40 hours.

Projects

Tree Visibility in the Netherlands: A Viewshed‑Based Approach and Multiscale Geographically Weighted Regression Modeling Github link here

Developed an objective, data-driven strategy to compute tree visibility at a national level in the Netherlands using a novel application of the Viewshed algorithm. This approach not only allowed for the generation of a nationwide tree visibility dataset but also facilitated an in-depth analysis of tree visibility inequality using the weighted GINI index. Furthermore, spatial machine learning was employed to model and explore associations between tree visibility and selected socio-economic variables. The proposed method is reproducible and provides a comprehensive framework for assessing and understanding the spatial patterns of tree visibility at a national scale.

Gender Representation in English Songs (1960‑2020)

This project analyzes gender representation in English song lyrics from 1960 to 2020. Using machine learning and semantic analysis, we examined thematic patterns, word associations, and emotional nuances (VAD scores) to uncover biases and differences in how male and female artists portray gender. Our findings reveal significant insights into gender representation in popular music over time.

Selenium Real Estate Scraper for the Slovak Market Github link here

This project was developed in response to a high demand for a comprehensive real estate scraper for the Slovak market. The project leverages Selenium and BeautifulSoup to extract property data from nehnutelnosti.sk, a leading real estate platform in Slovakia. This scraper fills a significant gap in the market, providing users with a valuable tool for gathering property data that was previously unavailable.

The Edinburgh University Guardian Ranking Dashboard Tableau link here

The Edinburgh University Guardian Ranking Dashboard is a Tableau-based tool that visualizes the university’s performance in the Guardian University Rankings. It allows users to explore trends in overall ranking, student satisfaction, subject rankings, and financial metrics, while also providing comparisons to other universities. I developed this dashboard for university stakeholders, prospective students to better understand the university’s position in the UK higher education landscape.

Current Readings

Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems by Aurélien Géron