Tadej Štajner
Contact
tadej@tdj.si
Things I do
- Verticals: natural language processing, conversational systems, speech recognition, mobility and traffic modeling, sensor fusion, process mining, multilingual search
- Machine learning and data science:statistical modeling, anomaly detection, knowledge management, information retrieval, semantic web, machine learning lifecycle, neural network architectures, model tuning and search, patent creation
- Software engineering: data engineering, numerical computing, distributed systems, high performance computing, SQL, web development, concurrent development, microservices, systems programming, DevOps, privacy engineering, real-time systems
- Product: experiment design, product validation, prototyping
Current position
Open for new opportunities.
- Senior Machine Learning Engineer, SumUp, Berlin, Germany ((November 2020 - current)
- building a machine learning and data platform for risk management use cases using infrastructure-as-code, MLOps, platform intelligence
Previous positions
- Lead Data Scientist, HERE Technologies, Berlin, Germany (March 2015 - March 2020)
- Developing machine learning solutions for automotive products and services
including vehicle destination prediction, vehicle departure time prediction, traffic-aware travel time prediction, natural language understanding for voice input, high-precision vehicle positioning, forecasting and simulation systems
- Developing digital twin simulations for forecasting service behavior to support product decisions
- Developing distributed data processing software facilitating the above goals
- Developing, deploying, maintaining automated anomaly monitoring systems for the data quality of automotive safety services
- Mentoring MSc students and junior developers
- Data engineer, TVBeat, Ljubljana, Slovenia
- Processing and analysis of real-time TV viewer habits
- Co-founder, lead developer, Magazinius, Ljubljana, Slovenia (May 2013 - April 2014)
- A next-gen digital publishing and design tool with responsive automated layout.
- Research assistant, Artificial Intelligence Laboratory, Jožef Stefan Institute, Ljubljana, Slovenia (December 2009 - April 2014)
-
- Named entity disambiguation using background knowledge (linking text with knowledge - demo)
- Triple store for fast entity retrieval and entity graph algorithms (knowledge graph database - demo)
- Named entity extractor for slovene (finding mentions of proper names, part of enrycher, available here)
- Summarizing microposts in social media (Twitter summarization - show only most interesting posts)
- Multi-lingual sentiment analysis for English and Spanish
- Fast training of sentiment analysis models using active learning (demo)
- Recommender system for contextualization of knowledge work (Suggest documents and e-mails related to what you're currently doing)
- Developer, Zemanta (May 2008 - November 2008)
- Content analysis back-end development
- NLP, Python, C++, Linux
- Developer, SRC.SI (June 2004 - September 2007)
- Development of applications for business process support (Struts/Java)
- Front-end development, UI (HTML, JavaScript, JSP)
Presence
Other activities
- Yahoo! Research, Barcelona: Visiting Researcher: June 2011 - October 2011 (Research on summarization models for social media content in response to news)
- W3C Working group: Multilingual Web - LT (development of a standard for the process of localization and internationalization; my focus is on integration of automated language tools in the content authoring and enrichment step)
Education
Selected publications and patents
- Cross-lingual document similarity estimation and dictionary generation with comparable corpora: Štajner T, Mladenić D, Knowledge and Information Systems (2018)
- Automatic selection of social media responses to news: Štajner T, Thomee B, Popescu A, Pennacchiotti M, Jaimes A, Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2013)
- Entity resolution in texts using statistical learning and ontologies
Štajner T, Mladenić D, In Proceedings of the Asian Semantic Web Conference (2009)
- Eduardo Vellasques, Tadej Štajner, Augusto Hentz, Olivier Dousse: Method and apparatus for providing mobility-based language model adaptation for navigational speech interfaces, Patent US10670415B2, 2017
- Tadej Štajner, Olivier Dousse, Eduardo Vellasques, Augusto Hentz: Method and apparatus for providing global voice-based entry of geographic information in a device, Patent US10249298B2, 2017
- Tadej Stajner, Olivier Dousse, Daniel Seyde, Dmitry Skripin: Apparatus and associated method for use in updating map data, Patent US20180224285A1, 2017
More:
Presentations