Amsterdam, North Holland, Netherlands
Applied ML Researcher at eBay, working on large-scale machine learning systems for e-commerce, including deep learning models, retrieval architectures, and production-grade pipelines. My work spans research, experimentation, and engineering, bridging the gap between ML science and high-performance software. I’m a graduate of UC Berkeley and École Centrale Paris, with a strong foundation in machine learning, NLP, statistics, and software engineering. I enjoy tackling challenging modeling problems and bringing ML solutions all the way from prototype to production.
Working as an applied researcher within eBay seller guidance team, with a focus on Deep Learning, Generative AI and semantic search algorithms.
Worked on various projects within the data science team, including: - Implementing a DeepFM model pipeline for complementary products recommendations - Designing and maintaining CI/CD pipelines with Jenkins, Airflow workflows, and Kubernetes orchestration for scalable deployment - Working on a text-to-SQL tool leveraging LLMs to translate natural language questions into SQL queries with awareness of table schema
- Trained and deployed a Transformer-based model to categorize Alexa search queries between Amazon stores - Created a new candidates generation system using factorization machines to surface relevant products to users - Researched and engineered new features for a personalized ranker model for products recommendations
- Implemented and trained a Factorization Machine (FM) model for personalized property ranking, leveraging historical booking behavior and contextual user features. - Designed and deployed a multi-armed bandit algorithm to dynamically optimize notification send times in the mobile app.
ML Modelling, insights and inference through various projects including: - Predictive modelling - Deep-dive researches and analyses - A/B testing - Automation of KPI monitoring and reporting