Computer Vision and Machine Learning Specialist
3DUniversum
Nov 2021 – Aug 2022 Amsterdam, Netherlands
- Researched a novel approach for a deepfake model by combining the state‐of‐the‐art in one‐shot face swap models (CNNs) with the attention in Transformer Vision models (ViT). The model is capable of swapping the faces of two individuals in a variety of situations.
Data Science Research Intern
Visma Connect V.B.
August 2018 – August 2020 Heerlen/The Hague, Netherlands
- Performed research to evaluate and mitigate gender bias in the Dutch contextualized word embeddings like BERTje, and analyzed the performance in downstream tasks leading to the publication of my thesis.
- Developed a data capture solution that uses optical character and named entity recognition technologies to automate the retrieval of key data from contracts, removing the need for people to manually enter them into the system. (OCR & NER).
- Analyzed the data from an external client and developed a model to forecast tenant departure date. I was responsible for feature selection, model validation, and hyperparameter tuning using gradient boosting trees.
Data Engineer Intern
Blockformer
June 2018 – August 2018 Maastricht, Netherlands
- Created a data pipeline to automate the data collection across multiple online sources using Python, JavaScript, and SQL.
- Identified new external data sources to supplement the company data.
- Effectively mined unstructured data and categorized it using sentiment analysis.
