Actively mentoring students in AI, machine learning, and Industry 4.0, I'm dedicated to sharing my expertise and nurturing the next generation of innovators, helping them excel in these rapidly evolving fields.

Working on the development of cutting-edge techniques and tools for handling extensive databases of unstructured and streaming data, my work encompasses cross-lingual news recommendation, web mining, event detection, and event extraction. As a dedicated contributor to the QMiner open-source project, I strived to advance the capabilities of these tools and techniques, paving the way for more efficient and effective data.

Worked on the development of innovative techniques and methodologies for semi-automatic ontology construction. I explored the extraction of insights from diverse unstructured data sources, including text, graphs, and images. My research focused on advancing visualization approaches for large text corpora, text streams, and ontologies, as well as cross-lingual information access and classification strategies.

Concentrating on optimization within the area of Text Mining, I dedicated my efforts to implementing state-of-the-art SVM models specifically designed for sparse vectors. This focus allowed me to delve into the nuances of this approach and contribute to the advancement of efficient, high-performance solutions for large-scale textual data analysis.

Leading a dynamic company focused on revolutionizing the manufacturing sector through advanced AI solutions. As the driving force behind our product's development from inception to market, I've successfully guided a talented team in creating a competitive edge within the industry landscape. We managed to delivered innovative and intelligent solutions tailored to our clients' needs.

Led a team of dedicated consultants in implementing text analysis, information extraction, and data mining projects into production, we focused on: (a) employing a fusion of deep-learning and semantic approaches to extract structured information from emails, (b) tackling anomaly detection and root-cause analysis within machine logs of large computer clusters, and (c) developing systems to support media ad teams with valuable audience insights.

Core contributor to the database engine and data scientist, I have played a significant role in enhancing event extraction algorithms, utilizing my expertise to optimize data processing and analysis.

As a Marie Curie Postdoctoral Fellow, I engaged in a comprehensive study of large-scale financial networks by leveraging privileged central bank datasets. My research aimed to predict corporate defaults and assess macroeconomic conditions at the country level. In parallel, I explored the application of deep learning methodologies to perform aspect-based sentiment analysis on text data sourced from social media platforms.

As a Research Engineer, I dedicated my efforts to creating cutting-edge solutions for web analytics and real-time news recommendation systems. My role encompassed providing valuable consultation on the practical application of text and data mining, machine learning, and artificial intelligence within industrial settings, ensuring optimal results and efficiency.

As a Postdoctoral Researcher, I focused on advancing techniques for event detection and extraction by employing active learning and deep learning methodologies. In addition to contributing to the management and mentorship of PhD students, I also actively participated in preparing research grant proposals.

Worked on the LarKC project, where my primary focus was on the seamless integration of text mining and machine learning techniques with the Cyc ontology and associated tools. This interdisciplinary approach aimed to enhance knowledge representation and reasoning capabilities for more effective outcomes

As an intern with the Web and Multimedia group, I led the design and development of a real-time news recommendation engine that significantly enhanced personalization on Bloomberg.com. This scalable system outperformed existing state-of-the-art solutions by innovatively utilizing unstructured information, all while maintaining real-time performance on a website with tens of millions of daily page views. I was actively involved in the entire process, from research to the implementation of the production system.

During a 6-month research internship with the Integrated Systems group at MSR Cambridge, I concentrated on discourse analysis within internet forums. My work involved identifying and quantifying discourse quality versus off-topic and hate speech, pinpointing knowledgeable and influential members in technical Q&A forums, and enhancing information retrieval in these forums. Additionally, I developed methods for the automatic analysis of website navigational structures, including automatic extraction of sitemaps from large websites, and visualization of extensive websites based on their navigation structures and web traffic.

Working on algorithms for multilingual text mining and information retrieval. Implementation and experimentation using (a) canonical correlation analysis for cross-lingual information retrieval and classification, and (b) string kernels for classification of texts in languages with many inflections (e.g. Slovene).