Main focus and expertise on deep-learning (representation-learning and generative-models), NLP, computer-vision, statistics and data-science (data mining, analysis, prediction and visualization). [Python is love]
Currently data-scientist at Zalando Dublin. Working on the design, development and deployment of machine-learning solutions in production. Manly relying on Python, Kubernetes, Tensorflow and AWS.
Projects I’ve worked on at Zalando include:
- named-entity-recognition and entity-linking surfaced as Scala APIs and enrichment job running on Spark.
- color-similarity API in Python to be used for A/B testing on Zalando website product page.
- pipeline to tag products based on image data. Automated dataset collection via scraping and crowsourced annotations. Trained deep-learning models for binary and multiclass classification. Served tags via Tensorflow Serving in event-driven architecture.
- embeddings + shallow classifiers for products classification. Compare SOTA models like StyleGAN, CLIP, DINO with in-house representations.
- synthetic-data-generation for products classification, relying on generative-models and computer-graphics pipelines.
Previously at IBM Watson, where I worked mainly with conversational-agents (chatbots), information-retrieval systems and visual-recognition, for several external clients.
EIT alumni, with a double master-of-science degree from ELTE and University of Trento, distributed-service-systems specialization.