At Legit, I’m researching and developing a Personal Data Search module, a web application that uses semantic search. I work with Hugging Face and Sentence Transformers for text embeddings and retrieval, alongside Apache Solr for indexing and search.
At Profico, I developed a customer churn prediction model, from data preparation and feature engineering through evaluation and explainability. I also delivered product-driven analysis, visualizations, and reporting for multiple clients.
At Profico, I designed and maintained data pipelines, ETL processes, and data warehousing using SQL and BigQuery. I also set up web and product analytics tracking and developed dashboards and reporting solutions.
At Faculty of Science, University of Split, I taught data science, machine learning, neural network architectures, and artificial intelligence. My other experience includes maintaining distributed database integration tests at Ericsson Nikola Tesla and analyzing spatio-temporal weather data in Python.
