Pavel Kordík

AI Researcher | CEO at Recombee | Bridging Academia & Industry

Making the world a better place by supporting others in what they do best. Creating teams that drive change across industry, academia, and non-profit sectors. Developing scalable, AI-based digital products that disrupt existing problems in education, academia-industry collaboration, recommender systems, personalized search and beyond. Passionate about research-based approaches that lead to superior solutions, while championing openness, knowledge sharing, and diverse international teams where motivated individuals empower each other.

Pavel Kordík

Business Impact

Recombee

CEO and Co-Founder of Recombee, delivering cutting-edge AI-powered recommender systems as a service. Supporting Tomáš Řehořek and other co-founders in the mission to bootstrap a company that now serves hundreds of millions of active users worldwide. Recombee has grown to more than 50 employees with ARR approaching 10M EUR.

Startup Ecosystem

Improving how startups spin out from academia (via CapTTict project), advising multiple early-stage startups on AI implementation, and consulting with Tensor Ventures on evaluating AI maturity of potential investments.

experts.ai

Co-founder of experts.ai, a platform connecting AI expertise with industry needs, fostering collaboration between academic experts and businesses in the field of artificial intelligence. The platform enables research teams to showcase their achievements and success stories, licenses, and offer research services to industry. Departments, faculties, and universities can enhance their websites with experts.ai widgets to be more attractive for business partnerships.

edumatch.ai

Co-founder of edumatch.ai, an innovative digital platform leveraging recommender systems and AI agents to match students with real-world opportunities in companies. Edumatch offers universities a robust platform for digital partnerships while enabling companies to access talented students across multiple universities simultaneously.

Bridging Academia & Industry

Dedicated to transforming the relationship between academic excellence and commercial innovation in AI.

Vision

My core mission is fostering synergistic relationships between academic research and industry applications. I believe the most impactful innovations emerge when cutting-edge research meets real-world challenges.

Academic to Industry

  • Translating research breakthroughs into scalable AI solutions
  • Guiding companies in adopting state-of-the-art AI technologies
  • Connecting talented researchers and students with industry opportunities
  • Developing educational pathways aligned with industry needs

Industry to Academia

  • Bringing real-world challenges to academic research
  • Providing industry datasets for academic investigation
  • Supporting AI education through industry partnerships
  • Mentoring students in solving practical AI problems

Research & Academia

Research Focus

Leading research in Artificial Intelligence, Machine Learning, and Recommender Systems at Czech Technical University in Prague.

Research Leadership

  • Coordinator of the Computational Intelligence Research Group
  • Active member of the Data Science Laboratory and Recombee Research Lab
  • Principal Investigator of multiple research projects
  • Supervising several PhD and industrial PhD students.
  • Motivating undergrad students to pursue their own research interests

Key Research Areas:

  • Recommender Systems and Personalization
  • Machine Learning and Neural Networks
  • Meta-learning and Transfer Learning
  • Computational Intelligence and Optimization
  • Predictive Modeling and Auto ML approaches
  • Information Visualization and Knowledge Discovery
  • Data Mining, Signal Processing and Clustering

Research Impact

My research bridges theoretical advancements with practical applications, particularly in recommender systems. This synergy between academia and industry enables the rapid transfer of cutting-edge algorithms into real-world solutions that serve millions of users worldwide.

EU Research Projects

Leading innovative research initiatives funded by European Union programs as Principal Investigator.

PoliRuralPLUS

Period: January 2024 – December 2026

Call: HORIZON-CL6-2023-COMMUNITIES-01-2

Project Focus

Fostering Sustainable, Balanced, Equitable, Place-based and Inclusive Development of Rural-Urban Communities Using Specific Spatial Enhanced Attractiveness Mapping ToolBox

Rural Development Spatial Mapping Sustainability

FOCAL

Period: November 2024 – November 2027

Call: HORIZON-RIA (Horizon Europe Research and Innovation Actions)

Project Focus

Developing a platform for climate change impact understanding, enabling better climate adaptation strategies through innovative data analysis and visualization techniques.

Climate Change Data Platform Impact Analysis

CapTTict

Period: January 2024 – June 2026

Call: Interreg Danube Region Programme

Project Focus

Strengthening Capacities for Technology Transfer and ICT application in the Danube Region Agri-food sector, bridging innovation gaps between research and industry implementation.

Technology Transfer Startups EDIHs, experts.ai

Non-Profit Initiatives

prg.ai

Co-founder and active board member of prg.ai, a non-profit initiative aimed at transforming Prague into a European AI hub. Working to connect academia, industry, and the public sector to foster innovation and education in artificial intelligence.

  • Strategic planning for AI ecosystem development (focusing on three key areas: Ecosystem, People, and Impact)
  • Coordination between universities, companies, and public institutions
  • Supporting AI education and research initiatives

aidetem.cz

Co-founder and technical advisor at aidetem.cz, a non-profit organization dedicated to introducing children to artificial intelligence in an engaging and responsible way.

  • Technical guidance for the Tiny application development
  • AI Research and development coordination
  • Strategic planning for AI education programs
  • Ensuring responsible AI practices in children's education

Professional Journey

A visualization of my parallel paths in academia and industry.

  • 2007

    Ph.D. in Artificial Intelligence

    Completed Ph.D. at Czech Technical University in Prague with a focus on meta-learning and neural networks optimization.

  • 2015

    Co-founded Recombee

    Co-founded Recombee to revolutionize recommender systems as a service, translating academic research into scalable industry solutions.

  • 2016-2024

    Vice Dean for Development at FIT CTU

    Appointed Vice Dean at the Faculty of Information Technology, focusing on transforming university-industry collaborations and implementing the partnership program for companies.

  • 2017

    Co-founded experts.ai

    Established experts.ai as a platform connecting AI expertise with industry needs, fostering collaboration between academic experts and businesses in artificial intelligence. The platform enables research teams to showcase their achievements and offer research services to industry, while universities can enhance their websites with experts.ai widgets.

  • 2018

    Associate Professor at CTU in Prague

    Appointed Associate Professor at the Czech Technical University in Prague, Faculty of Information Technology, continuing to advance academic research and education in AI and machine learning.

  • 2018

    Recombee Scale-up

    Contributed to Recombee's expansion to serve tens of millions of users worldwide, developing cutting-edge AI-powered recommendation engines.

  • 2019

    Co-founded prg.ai

    Co-initiated prg.ai to transform Prague into a European AI hub, connecting academia, industry, and public sector for innovation and education in AI.

  • 2021

    Co-founded aidetem.cz

    Established aidetem.cz to introduce children to AI in a responsible and engaging way, fostering future AI literacy and ethical understanding.

  • 2024

    Co-founded edumatch.ai

    Launched edumatch.ai, student centric platform leveraging AI to run digital partnerships between universities and companies and match students with real-world opportunities.

  • 2025

    RecSys25 General Co-Chair

    Leading the organization of the 19th ACM Conference on Recommender Systems (RecSys) in Prague. With over 1,000 attendees from around the world, this major event brings together academic and industrial researchers to foster future developments and advancements in the field of recommender systems.

  • Present

Teaching

Academic Courses at CTU

Advanced Data Mining (ADM)

Advanced techniques in data mining, focusing on modern approaches to recommenders systems and search. Discussing theoretical concepts like bias variance decompositon or model confidence modelling.

Modern AI Architectures (MVI)

Exploring cutting-edge AI architectures and optimization strategies, bridging theoretical foundations with practical implementations.

AI Research Seminar

Leading discussions on current AI research trends, methodologies, and breakthrough papers, fostering critical thinking and research skills.

Student Opportunities

I'm actively supervising student projects, master theses, and PhD dissertations in the fields of AI, machine learning, and recommender systems. If you're interested in working on cutting-edge research with real-world applications, I invite you to explore the opportunities below.

Bachelor/Master Projects

Work on exciting projects involving machine learning and AI applications, often in collaboration with industry partners.

View Available Projects

Master Theses

Develop advanced solutions to challenging problems in AI, with potential for publication in top conferences and journals.

Explore Thesis Topics

PhD Opportunities

Join my research team to pursue cutting-edge research in recommender systems, deep learning, and AI architectures.

Discover PhD Topics

Interested in proposing your own topic? I'm always open to discussing novel ideas and research directions.

Contact Me Directly

Selected Publications

A curated selection of my most impactful research publications across key areas of AI and recommender systems.

Author of 90+ research publications with 500+ citations. Complete list available on Google Scholar and DBLP.

2024

beeFormer: Bridging the Gap Between Semantic and Interaction Similarity in Recommender Systems

Vančura, V., Kordík, P., & Straka, M.

Proceedings of the 18th ACM Conference on Recommender Systems (2024)

Improve recommendation of cold start items by training transformers on interactions.

2023

Overcoming the Cold-Start Problem in Recommendation Systems with Ontologies and Knowledge Graphs 4+ citations

Kuznetsov, S., & Kordík, P.

European Conference on Advances in Databases and Information Systems, pp. 591-603 (2023)

Novel approach to addressing the cold-start problem in recommendation systems using semantic technologies.

Personalised Recommendations and Profile Based Re-ranking Improve Distribution of Student Opportunities

Žid, Č., Kordík, P., & Kuznetsov, S.

Computational Intelligence in Security for Information Systems Conference (2023)

Improving educational opportunities distribution through personalized recommendation systems.

Bridging Offline-Online Evaluation with a Time-dependent and Popularity Bias-free Offline Metric for Recommenders 5+ citations Best Paper Award

Kasalický, P., Alves, R., & Kordík, P.

2nd Edition of EvalRS: A Rounded Evaluation of Recommender Systems - KDD Workshop, Long Beach, CA, USA (2023)

Novel evaluation metric addressing the gap between offline and online recommender system performance. Best Paper Award at EvalRS 2023.

Decorelated Weight Initialization by Backpropagation

Kovalenko, A., & Kordík, P.

International Conference on Artificial Neural Networks, pp. 546-550 (2023)

Novel approach to neural network weight initialization improving convergence and stability.

2022

Scalable linear shallow autoencoder for collaborative filtering 14+ citations

Vančura, V., Alves, R., Kasalický, P., & Kordík, P.

Proceedings of the 16th ACM Conference on Recommender Systems, pp. 604-609 (2022)

Highly efficient autoencoder approach for large-scale collaborative filtering applications.

Adapting the Size of Artificial Neural Networks Using Dynamic Auto-Sizing 1+ citation

Cahlík, V., Kordík, P., & Čepek, M.

IEEE International Conference on Computer Sciences and Information Technologies (2022)

Innovative approach to dynamically adjusting neural network architecture during training.

Learning to Optimize with Dynamic Mode Decomposition 4+ citations

Šimánek, P., Vašata, D., & Kordík, P.

International Joint Conference on Neural Networks, pp. 1-8 (2022)

Applying dynamic mode decomposition to optimization problems for faster convergence.

RepSys: Framework for Interactive Evaluation of Recommender Systems 5+ citations

Šafařík, J., Vančura, V., & Kordík, P.

Proceedings of the 16th ACM Conference on Recommender Systems, pp. 636-639 (2022)

Interactive framework for comprehensive evaluation of recommender systems, enabling interactivity and visual inspection.

2021

Transfer learning based few-shot classification using optimal transport mapping from preprocessed latent space of backbone neural network 22+ citations 2nd Place Award

Chobola, T., Vašata, D., & Kordík, P.

AAAI Workshop on Meta-Learning and MetaDL Challenge, pp. 29-37 (2021)

Innovative approach to few-shot learning using optimal transport mapping for enhanced performance. Won 2nd place in the Meta-Learning challenge competition.

Deep Variational Autoencoder with Shallow Parallel Path for Top-N Recommendation (VASP) 17+ citations

Vančura, V., & Kordík, P.

International Conference on Artificial Neural Networks, pp. 138-149 (2021)

Hybrid deep-shallow neural architecture for improved recommendation accuracy and efficiency.

Stratified Cross-Validation on Multiple Columns 9+ citations

Motl, J., & Kordík, P.

IEEE 33rd International Conference on Tools with Artificial Intelligence, pp. 26-31 (2021)

Enhanced cross-validation technique that preserves multiple stratification constraints.

Novel Data Mining-based Age-at-death Estimation Model using Adult Pubic Symphysis 3D Scans 2+ citations

Buk, Z., Štepanovský, M., Koterová, A., Velemínská, J., Brůžek, J., & Kordík, P.

Information Technologies – Applications and Theory, pp. 46-52 (2021)

Applying data mining to forensic anthropology for improved age-at-death estimation.

2019

Chameleon 2: an improved graph-based clustering algorithm 35+ citations

Barton, T., Bruna, T., & Kordik, P.

ACM Transactions on Knowledge Discovery from Data (TKDD), Vol. 13(1), pp. 1-27 (2019)

Enhanced graph-based clustering algorithm with improved performance for complex data structures.

2018

Discovering predictive ensembles for transfer learning and meta-learning 41+ citations

Kordík, P., Černý, J., & Frýda, T.

Machine Learning, Vol. 107(1), pp. 177-207 (2018)

Advanced ensemble methods for improving transfer learning and meta-learning capabilities in ML systems.

2016

Neural turing machine for sequential learning of human mobility patterns 36+ citations

Tkačík, J., & Kordík, P.

International Joint Conference on Neural Networks (IJCNN), pp. 2790-2797 (2016)

Application of Neural Turing Machines to predict human mobility patterns with enhanced accuracy.

Reducing cold start problems in educational recommender systems 20+ citations

Kuznetsov, S., Kordík, P., Řehořek, T., Dvořák, J., & Kroha, P.

International Joint Conference on Neural Networks (IJCNN), pp. 3143-3149 (2016)

Novel methods to address cold start challenges in educational recommender systems for improved learning experiences.

2010-2012

Selecting representative data sets 145+ citations

Borovicka, T., Jirina, M. Jr., Kordík, P., & Jirina, M.

Advances in data mining knowledge discovery and applications 12, 43-70 (2012)

A fundamental work on optimizing data selection for machine learning models, significantly improving model generalization.

The age at death assessment in a multi-ethnic sample of pelvic bones using nature-inspired data mining methods 62+ citations

Buk, Z., Kordik, P., Bruzek, J., Schmitt, A., & Snorek, M.

Forensic Science International 220 (1-3), 294.e1-294.e9 (2012)

Innovative application of data mining methods to forensic anthropology for improved age estimation.

Meta-learning approach to neural network optimization 67+ citations

Kordík, P., Koutník, J., Drchal, J., Kovářík, O., Čepek, M., & Šnorek, M.

Neural Networks, Vol. 23(4), pp. 568-582 (2010)

Novel approach to optimizing neural networks through meta-learning principles, enhancing performance and efficiency.

Contact

Get in Touch

For business inquiries, research collaboration, or academic matters, feel free to reach out.

kordik@gmail.com