Urszula Czerwinska

Senior Deep Learning engineer, Data Scientist, PhD

Download CV

Short Bio

Born in Poland and having gained diverse educational experiences across France and Singapore, I now proudly call Paris home. My journey, rooted in bio-mathematics, culminating in a Ph.D., has equipped me with profound expertise in complex systems, statistics, and programming, particularly in Python and R. These interdisciplinary experiences have sharpened my ability to excel in bridging the gap between theoretical knowledge and practical application, specially in the realms of deep learning and data science.

As a Senior Deep Learning Engineer - Data Scientist at Adevinta in Paris, my current focus is on spearheading advancements in computer vision and multimodal AI models. I lead the strategic development of foundation models that drive product innovation and enhance e-commerce solutions. My work leverages cutting-edge technologies like PyTorch, TensorFlow, and Kubernetes, where I guide my team to deliver exceptional results. Notably, I’ve improved scene-text extraction capabilities by 33% and optimized multi-class classification models to increase accuracy by 45%.

My role at Adevinta also encompasses leadership beyond technical accomplishments. As the head of the Adevinta ML Guild, I facilitate paper reading sessions and am actively involved in recruiting a Ph.D. expert to further strengthen our team. This position allows me to nurture a culture of continuous learning and innovation, a pursuit I am deeply passionate about.

Before my role at Adevinta, I served as a Senior Data Scientist and NLP Engineer at the French Supreme Court, where I developed language models and named entity recognition systems. These innovations automated the anonymization of judicial documents, boosting productivity by 45%. My role required close collaboration with academia, public institutions, and the European Commission was a significant part of this role, underscoring my ability to integrate research with practical solutions.

I am also an advocate for diversity in tech. My previous role as an ambassador for the Pivigo Data Science hub and my involvement with the Women in Healthcare and Data Science Paris meetup, where I organized events bridging academia and industry, reflect this commitment.

In addition to my professional roles, I’ve contributed a chapter on ML model interpretability to Applied Data Science in Tourism a Data Science book and have shared my work at various international conferences and through publications.

I’m driven by the opportunity to build impactful AI models that solve real-world problems, foster a culture of continuous learning, and promote the responsible use of AI. You can bump into me while attending some of AI, Deep Learning conferences such as ICCV, ECC, ICR or NeurIPS. Let’s connect to explore how we can collaborate to drive innovation and harness the power of AI to achieve transformative results.

Professional experience

I have been developing diverse projects in Computer Vision, NLP, Data Science, bioinformatics, biology and others. Have a look at my Works .

Education

Academic

For publications see my orcid profile, for posters, science awards and talks please see my academic CV .