The 10th international conference on Machine Learning and Artificial Intelligence applications,
taking place in person in Prague and online.

Machine Learning Prague 2025

– , 2025

Registration

World class expertise and practical content packed in 3 days!

You can look forward to an excellent lineup of 40 international experts in ML and AI business and academic applications at ML Prague 2025. They will present advanced practical talks, hands-on workshops, and other forms of interactive content to you.

What to expect

  • 1000+ Attendees
  • 3 Days
  • 40 Speakers
  • 10 Workshops

Phenomenal Confirmed speakers

Hava Siegelmann

Provost Professor, University of Massachusetts Amherst

Senior faculty for bio-inspired AI, Dr. Siegelmann is an internationally known UMass Provost Professor in Computer Science and a recognized expert in neural networks. She is a core member of the University of Massachusetts Neuroscience and Behavior Program and director of the Biologically Inspired Neural and Dynamical Systems (BINDS) Laboratory. She has been particularly acclaimed for her groundbreaking work in computation beyond the Turing limit, and for achieving advanced learning capabilities through a new type of Artificial Intelligence: Lifelong Learning. Siegelmann conducts highly interdisciplinary research in next-generation machine learning, neural networks, intelligent machine-human collaboration, and computational studies of the brain - with application to AI, data science, and high-tech industry. Prof. Siegelmann is a co-inventor of the Support Vector Clustering (SVC) algorithm, which is widely used across industry and government. Among her recent Nature publications is Biological Underpinning of Lifelong Learning AI, a bio-inspired replay algorithm for advanced lifelong learning, dual fractal structure & function of the human brain, and identification of a previously unknown brain connectome mechanism, which enables cognitive abstraction.

Stanislav Fort

Senior Research Scientist, Google DeepMind

Ondřej Filip

Research Team Lead, Seznam.cz

Ondřej's current mission is training Czech LLMs at Seznam.cz. At the same company, he gained extensive experience prototyping machine learning models for various search verticals, including web, images, and maps. He holds a degree in Artificial Intelligence from Charles University in Prague.

Alexandr Vendl

Senior Data Scientist, Dataclair

Since joining Dataclair in 2022, Alexandr Vendl has been applying his expertise as a Senior Data Scientist to address complex business challenges. His primary focus lies in the application of Large Language Models (LLMs) and the development of fraud detection solutions. A graduate of Czech Technical University, Alexandr holds a Master's degree in Laser Physics. His academic background in mathematical modeling and the use of deep learning in femtosecond laser physics has provided him with a unique perspective when approaching AI and data science projects.

Jakub Sochor

Chief Technology Officer, Innovatrics

Jakub Sochor is the Chief Technology Officer at Innovatrics, a world-class provider of biometric technology and solutions, consistently ranked in top positions by NIST. Before joining Innovatrics, Jakub worked at Google, contributing to the development of Google Assistant. He holds a PhD from Brno University of Technology, specializing in computer science and artificial intelligence.

Philipp Wendland

AI Consultant, Deloitte Consulting

Philipp Wendland is a Senior Consultant in the Deloitte AI Institute, Germany. He is an expert in AI, with a focus on particularly Generative AI. His approach to AI is multifaceted ranging from co-authoring thought-leadership pieces, conducting workshops and C-level briefings to implementations of (Generative) AI applications. Making AI approachable and the inherit complexities understandable is one of his key undertakings. In this effort he has trained small groups of executives on AI as well as over 1,500 employees on e.g., Prompt Engineering techniques. Philipp holds a bachelor’s degree in physics from the university of Heidelberg and a masters in “Robotics, Cognition and Intelligence” from the Technical University of Munich.

Jakub Tomasz Gnyp

Computational Scientist, International Centre for Theory of Quantum Technologies

Jakub Tomasz Gnyp is a lab technician with tenure at the Condensed Matter Spectroscopy Division of the University of Gdańsk. Since 2022, he has also conducted research on quantum key distribution optimization at the International Centre for Theory of Quantum Technologies (ICTQT), as part of the Quantum Cybersecurity and Communications Group. He has contributed to projects such as "Theoretical Study of Time-Bin Entanglement Properties for Quantum Internet" and "Development of a Quantum Repeater in Optical Fiber Networks for Quantum Internet," in collaboration with the Electronics and Telecommunications Research Institute (ETRI) in South Korea.

His current research focuses on the applications of computational intelligence, machine learning, and stochastic analysis in physics. He is particularly interested in QKD error correction algorithms, near-infrared spectroscopy, and modeling weather at sea.

Jakub is a mathematician and physicist with data science experience. He worked on offshore wind farm construction projects in Antwerp, Belgium, with Ultra NDT, and has completed various assignments for companies in Poland, including Jubitom, QCG, StatXplorer, 7willows, and Talkersi. Currently, as an employee of the University of Gdańsk, he is continuing his education there.

Tobias Kietreiber

Junior Researcher, St. Pölten University of Applied Sciences

Tobias Kietreiber is a Junior Researcher at the University of Applied Sciences St. Pölten, specializing in reinforcement learning, imitation learning, and explainable AI. With a strong foundation in mathematics, Tobias explores practical applications of these techniques, including hate speech detection and user group imitation on websites. His work involves developing intelligent systems that not only learn complex tasks through demonstration but also identify harmful online content and mimic user behaviors to enhance website interaction. Tobias also focuses on explainable AI, ensuring transparency and interpretability in machine learning models, contributing to ethical and effective AI development.

Sebastian Eresheim

Junior Researcher, St. Pölten University of Applied Sciences

Sebastian Eresheim is a Junior Researcher at University of Applied Sciences St. Pölten. He has a master's degree in Information Security from UAS St. Pölten, but also studied Technical Mathematics at Technical University Vienna. His research interests are in Reinforcement Learning and its real-world applications. His work involves finding optimal defense strategies in cyber systems to prevent and/or mitigate cyber threats.

Alexander Buchelt

Junior Researcher, St. Pölten University of Applied Sciences

Alexander Buchelt is a Junior Researcher at University of Applied Sciences St. Pölten and a PhD Student at the University for Natural Resources and Life Sciences Vienna, specializing in artificial intelligence and digital twins. His current research revolves around developing advanced AI algorithms for autonomous drones, aiming to optimize their application in forestry management to improve data collection and analysis in forest environments. His work is at the forefront of integrating AI-driven solutions into sustainable natural resource management.

Filip Roskovec

Head of Data Science, O2/Dataclair

As a senior leader of the data science team at AI centre at O2 Czech Republic, Filip specializes in internal projects, including next best action prediction and developing machine learning models for mobile network optimization. He also organizes events to promote cross-team collaboration and communication, often focusing on the latest advancements in large language models (LLMs). Filip holds a Ph.D. in Numerical and Computational Mathematics from Charles University in Prague.

Stefan Josef

Senior Data Scientist, O2/Dataclair

Stefan is a Senior Data Scientist at Dataclair, AI centre at O2 Czech Republic, where he has been leading research efforts in applying advanced NLP and Deep Learning techniques to geospatial data. Drawing on years of experience in working with language models, Stefan recently shifted his focus on improving production RAG systems through a combination of synthetic data generation and model fine-tuning. Stefan holds a M.Sc. in Economics from Stockholm University, where he specialized in empirical macroeconomics.

Ondřej Finke

Senior Data Scientist, O2/Dataclair

 

Ondřej Finke is a Senior Data Scientist at Dataclair, where he has been working since 2023. His primary focus is on the practical application of Large Language Models in various projects. Ondřej holds a master’s degree in laser physics from Czech Technical University. Before joining Dataclair, he gained experience in experimental physics and data processing on large scale experiments, which has influenced his approach to problem-solving in AI and machine learning.

Kryštof Šaml

Researcher, Emplifi

Krystof Saml is a Researcher at Emplifi who develops practical AI solutions using data-centric approaches. He works on synthetic data generation systems and evaluation solutions for language models, focusing on improving data quality. Currently, he explores ways to combine LLMs with smaller specialized models to create more efficient AI systems.

Tomáš Sikora

Researcher, Emplifi

Tomas Sikora is a Researcher in Emplifi's Innovations department. He focuses on large language models (LLMs), agentic systems, and machine reasoning and their practical implementation through model training and tuning. He uses his expertise as part of the Search and Recommendation team, where he scales up solutions to address complex user requirements.

Jérémy Cochoy

CEO, Redstone Solution OÜ

Jérémy Cochoy is an expert in technology with a strong academic background. Holding a PhD in Computer Science and Mathematics with a focus on Persistent Homology, he leveraged his expertise to co-found Symphonia, an app that creatively transforms voices into music. Currently, as CEO of Redstone Solutions, Cochoy applies his skills in deep learning to the field of financial market forecasting. His career is a testament to the fusion of advanced scientific knowledge and practical technological applications, underscoring his commitment to driving innovation in complex fields. Beyond his professional realm, Cochoy's interests in music and other artistic pursuits reflect a multifaceted personality, equally engaged in intellectual and creative endeavors.

Szymon Bubak

CTO, Jiai

CTO at Jiai since 2024, focusing on Python, MLOps, and tech leadership. Founder leader of Silesia AI, fostering a strong AI community. I studied mathematics and spent nearly 8 years at Roche, where I led software development for drug discovery, improving processes across labs, enhancing developer efficiency, and integrating biotech tools critical for research on diseases like cancer and hemophilia. I also led rapid prototyping efforts at POC Now! in various industries.

Humera Noor Minhas

CTO, Digital Munich

Dr. Humera developed her love for data early on in her career and is still fascinated by its simplicity and power. As a PhD in Machine Learning, she has delivered several ML based projects successfully. Notably, she led the team that pioneered the use of machine learning for commercial-scale ad filtering. She is an entrepreneur and brings with herself twenty five years of professional experience with majors in computer vision and machine learning. Currently, she is the CTO of Digital Munich Tech GmbH, which is aimed to deliver Innovation as a Service and develop solutions for artificial intelligence, mixed reality, IOT, and data analytics. Dr. Humera loves to be in the nature and starts her day at sunrise by indulging in a book on a bench by the fields. Of course, with coffee! She can be reached at: https://www.linkedin.com/in/humeranoor/.

Varun Burde

Research Assistant, CIIRC, Czech Technical University

Varun Burde is a Ph.D. student at the Czech Technical University in Prague, where he is advised by Dr. Torsten Sattler and Dr. Pavel Burget. In addition to his doctoral studies, he serves as a Research Assistant at the CIIRC, CTU, within the Research Group at the Testbed for Industry 4.0. Varun holds a master’s degree in Cybernetics and Robotics from the same institution. His research focuses on advancing 3D reconstruction techniques to enhance robotic manipulation, aiming to push the boundaries of industrial automation. By leveraging cutting-edge computer vision algorithms, Varun seeks to drive innovation in automated systems, contributing to greater precision and efficiency in Industry 4.0 applications.

Tun Shwe

VP of Data, Quix

Tun is the VP of Data at Quix, where he leads data strategy and developer relations. He is focused on helping companies determine and implement their strategic data vision with real-time streaming data at the forefront. He was previously a Head of Data and Data Engineer at high growth startups and has spent his career leading T-shaped teams in developing data platforms and data-intensive AI applications. In his spare time Tun goes surfing, plays guitar, tends to his analogue cameras and nerds out on well designed things.

Tomáš Neubauer

CTO, Quix

Tomáš is Co-Founder and CTO at Quix, where he works as the technical authority for the engineering team and is responsible for the direction of the company across the full technical stack. He was previously technical lead at McLaren, where he led the architectural uplift of the real-time telemetry acquisition platform for the Formula 1 racing team. He later led platform development outside motorsport, reusing the know-how he gained from racing. In his spare time Tomáš likes to go mountain biking in the hills around Prague and explore the fine beers of Czechia.

Tomáš Tomeček

Senior Principal Software Engineer, Red Hat

With over a decade of experience as a Software Engineer at Red Hat, Tomas is passionate for leveraging technology to simplify life. He's focused on automation, integration, and solving complex problems. In the past year, Tomas works on using AI to boost efficiency and productivity. After work, he's a hiker, gardener and snowboarder.

Štefan Bunčiak

Principal Product Manager, Red Hat

With more than 10 years of experience in the software industry, Stefan Bunciak is currently the Product Manager for Red Hat Insights. He completed his master's degree in Informatics at Masaryk University in Brno and is skilled in project and people management, quality engineering, and software development. In his spare time, he plays violin in a folklore band.

Practical & Inspiring Program

Friday
Workshops

O2 Universum, Českomoravská 2345/17a, 190 00, Praha (workshops won't be streamed)

Registration –

Room D2 Room D3 Room D4 Room D6 Room D7
– coffee break
–

Utilizing Large Language Models for improved anti-tracking in web browsers

Room D2

Humera Noor Minhas, Digital Munich

Online tracking remains a significant privacy concern for internet users. Current solutions while effective have limitations in terms of coverage maintenance and precision. This workshop aims to leverage the power of LLMs to create a more robust adaptive and efficient anti-tracking system. We will explore the architecture of an LLM-based anti-tracking system developing the data pipeline and exploring how these models can be fine-tuned to analyze network requests page content and user interactions in real-time. The system's ability to understand the semantic context of web elements allows for more accurate identification of tracking attempts reducing false positives while improving detection rates of sophisticated trackers. A key focus will be on the practical challenges of implementing such a system within the constraints of a web browser environment. We'll discuss strategies for optimizing LLM inference to meet the real-time demands of browsing balancing accuracy with performance.

Accelerating AI Through Human Knowledge: Teaching to Imitate Experts and Win on the Race Track

Room D3

Alexander Buchelt, St. Pölten University of Applied Sciences
Sebastian Eresheim, St. Pölten University of Applied Sciences
Tobias Kietreiber, St. Pölten University of Applied Sciences

In this 3-hour hands-on workshop participants will explore the exciting world of Imitation Learning a powerful technique in artificial intelligence that allows agents to mimic expert behavior and excel in complex environments. Building on the fundamentals of Reinforcement Learning this workshop introduces the theory behind Imitation Learning and demonstrates how it can be applied to solve real-world problems efficiently. By guiding AI through expert demonstration imitation learning accelerates training especially in environments where traditional reinforcement learning might be time-consuming or difficult. Imitation Learning is crucial for AI systems that need to learn from limited data or human expertise such as autonomous driving robotics and gaming. In contrast to trial-and-error methods in reinforcement learning imitation learning allows models to replicate the strategies of experienced individuals drastically reducing training time and improving performance. Attendees will gain a deep understanding of how this approach combines the best of both supervised and reinforcement learning creating smarter faster decision-making systems.

Introduction to Algorithmic Trading: Hands-On Strategy Implementation with Real-World Data

Room D4

Szymon Bubak, Jiai
Jérémy Cochoy, Redstone Solution OÜ

Algorithmic trading has become a cornerstone of financial markets with automation and data-driven strategies driving the majority of transactions. This workshop provides a comprehensive hands-on introduction to the world of algorithmic trading aimed at students and professionals interested in financial markets who want to move beyond academic exercises and engage with real-world scenarios. Participants will develop a trading strategy using real-world data such as Bitcoin prices or stocks from the S&P 500 with the initial focus on achieving profitability under the assumption of no transaction fees. The 3-hour session will guide participants through the end-to-end process of designing implementing and backtesting a deep learning-based trading strategy. The workshop will be structured as follows: - Data Preparation: Participants will learn how to source and preprocess financial data preparing it for model input. - Feature Extraction: We will introduce simple features from the data to feed into a deep learning model. - Loss Function Design: The workshop will cover how to design a loss function tailored to trading strategies and objectives. - Model Training: Participants will implement and train a deep learning model using PyTorch. The workshop will focus on how to properly train models in low-data environments ensuring the model generalizes effectively by employing robust train/validation/test splits. Backtesting: We will backtest the model’s performance allowing participants to understand its strengths and limitations in various market conditions. Rather than dividing the session into theoretical and practical segments the entire workshop will seamlessly integrate theory and implementation. Participants will work on a live trading scenario continuously applying new knowledge as they progress through the workshop. By the end they will have created a complete training pipeline—from data preparation to model training and backtesting. A critical aspect of this workshop will be understanding the practical challenges of implementing algorithmic trading strategies in real-world markets. We will discuss the limitations of deep learning models when applied to financial data including how to mitigate overfitting in low-data environments. Additionally participants will explore the impact of transaction fees slippage and other market inefficiencies learning how these factors affect profitability. The workshop is designed for participants with a basic understanding of Python and machine learning but no prior experience with algorithmic trading is required. By the end of the session attendees will have gained practical skills in designing and implementing a trading strategy along with valuable insights into the intricacies of algorithmic trading in financial markets. Participants will be provided with all necessary code templates and data and they are expected to bring their own laptops to engage fully in the hands-on aspects of the workshop. This workshop offers a unique opportunity to bridge the gap between theoretical learning and real-world financial applications equipping participants with the tools and knowledge to pursue further exploration in algorithmic trading.

InstructLab: plug your knowledge into a model easily

Room D6

Štefan Bunčiak, Red Hat
Tomáš Tomeček, Red Hat

During this hands-on exercise you will learn what is InstructLab and how you can leverage it to easily extend Large Language Models with your data and run them on your infrastructure. The tool makes it easy to download run and chat with models locally on your laptop. InstructLab is a fully open-source project from Red Hat and the MIT-IBM Watson AI Lab that introduces Large-scale Alignment for chatBots (LAB). The paper behind it: https://arxiv.org/abs/2403.01081 The LAB method is driven by taxonomies which are largely created manually and with care. For a taxonomy you supply InstructLab then can generate synthetic data used to train a model. Everyone who has experience with LLMs can greatly benefit from this workshop. We will create our own knowledge documents use InstructLab to generate synthetic data out of them train a model from the data and chat with them.

Beyond Real-World Limitations - Mastering Synthetic Data Generation for Enhanced ML Performance

Room D7

Tomáš Sikora, Emplifi
Kryštof Šaml, Emplifi

As machine learning pushes into new frontiers the demand for diverse and representative datasets frequently outpaces the availability of real-world data. This workshop explores a spectrum of advanced techniques for synthetic data generation from traditional methods to cutting-edge AI agent-driven approaches. By embracing the principles of data-centric AI we'll start with traditional data generation techniques and progress to innovative AI agent-driven strategies. Throughout the workshop we'll demonstrate techniques that promise to overcome limitations in real-world datasets reshape the data landscape in ML and rigorously evaluate the quality of the generated data. Workshop Overview: - Setting the Stage: The Data Challenge in ML We'll begin with a brief historical context touching on pre-LLM approaches to synthetic data generation. This background will highlight the impact of recent advancements and set the stage for our deep dive into modern techniques. - Data-Centric AI: A Paradigm Shift We'll explore how the data-centric AI movement is reshaping our approach to machine learning. Participants will learn why focusing on data quality and targeted synthetic data generation can often yield better results than simply increasing dataset size or model complexity. -Evolution of Synthetic Data Generation with AI -LLM-Powered Data Creation: Leveraging large language models for synthetic data generation. -Multi-Agent Systems: Advancing to manually designed agent ecosystems for nuanced and accurate data production. -Automated Agent Workflows: Exploring the cutting edge with self-optimizing agent interactions for superior data quality. -Robust Evaluation in the Data-Centric Paradigm We'll emphasize rigorous evaluation techniques aligned with data-centric AI principles ensuring the effectiveness of synthetic data in real-world ML applications. This workshop is ideal for ML practitioners researchers and data scientists looking to overcome data quality and scarcity challenges. Participants should have a basic understanding of machine learning concepts and some experience with programming in Python. By the end of this workshop attendees will have a comprehensive understanding of how data-centric AI principles can be applied to synthetic data generation from LLM-based techniques to state-of-the-art agent-driven approaches. They'll be equipped with practical skills to implement these strategies potentially revolutionizing how they tackle data-related challenges in ML projects.

–
Lunch
– coffee break
–

3D reconstruction from Images and their application

Room D2

Varun Burde, CIIRC, Czech Technical University

This workshop will delve into recent advances in 3D computer vision and provide participants with practical hands-on experience in generating 3D reconstructions from image data. Attendees will explore cutting-edge techniques including neural radiance fields (NeRFs) Gaussian splatting multi-view stereo (MVS) and structure from motion (SfM) for surface reconstruction. The session will cover the fundamentals of 3D reconstruction focusing on how modern algorithms transform 2D images into detailed and accurate 3D models. Additionally participants will learn the essential steps of dataset creation and optimization for training advanced 3D reconstruction methods. A key feature of the workshop will involve capturing a set of images of objects and demonstrating how to systematically collect and organize data to ensure high-quality 3D model generation. Attendees will gain experience in building datasets tailored to different 3D reconstruction techniques such as NeRF and Gaussian splatting and optimizing them for improved accuracy and visual fidelity. This workshop is ideal for researchers engineers and enthusiasts seeking to understand the latest in 3D vision technologies with applications ranging from augmented reality and robotics to digital content creation.

Parallel Genetic Algorithms in Python

Room D3

Jakub Tomasz Gnyp, International Centre for Theory of Quantum Technologies

In this workshop we delve into the construction and implementation of parallel genetic algorithms (PGAs) using Python. Genetic algorithms (GAs) and evolutionary algorithms in general are powerful tools for solving optimization problems and when parallelized they offer significant speedups and efficiency improvements. Participants who learn PGAs will also be able to apply them in reinforcement learning. The workshop will have a limited amount of mathematics - instead the focus will be on both the idea behind PGAs and practical coding skills. Starting with the PyGAD library its uses and limitations will be discussed and presented with easy-to-understand examples. Later key aspects of parallel programming will be introduced such as recognizing CPU- and I/O-bound operations and the use of processes and threads respectively. Global lock in Python will be addressed as well as racing conditions. Therefore a basic understanding and implementation of locks barriers flags and shared memory in general will be achieved. To illustrate the practical applications of parallel genetic algorithms apart from minor examples the workshop features three major case studies. The first involves solving a labyrinth demonstrating how a parallel genetic algorithm can efficiently navigate complex search spaces and de facto interact with an environment. Participants will observe how the parallelization of GAs can lead to faster convergence on optimal paths compared to sequential approaches. Diversity in population will be addressed as well. The second case study explores the application of parallel genetic algorithms in quantum cryptography. In this domain GAs can optimize parameters for quantum key distribution protocols enhancing especially efficiency. By parallelizing the algorithm we can tackle the computational challenges of the vast solution spaces inherent in quantum cryptographic systems. The BB84 protocol will be the protocol in question explained without the quantum mechanics' mathematical rigor and the essentials of the protocol will already be implemented. The third and last case study will be a neural network in which hyperparameters will be optimized by a PGA in a Genetically Reinforced Learning scheme. Knowing how the PGA may interact with an environment and work on even very complicated functions this optimization task will be an easy step for those who have already seen the neural network. By the end of the workshop participants will have a solid understanding of how to implement and apply parallel genetic algorithms in Python with practical insights into their strengths and limitations. They will be equipped with the knowledge to extend these techniques to other domains fostering innovation in computational problem-solving.

Real-Time Anomaly Detection and Alerting in Financial Markets Using Stream Processing

Room D4

Tomáš Neubauer, Quix
Tun Shwe, Quix

In the world of financial markets the ability to detect and act on anomalies in real-time is crucial. This workshop will explore how to build a stream processing system that not only detects rapid changes in stock prices but also calculates key stock market indicators like the Relative Strength Index (RSI) Moving Average Convergence Divergence (MACD) or Bollinger Bands in real-time. Attendees will learn how to calculate these indicators in real-time to identify potential buy or sell signals and trigger instant alerts such as Slack messages to notify users of significant market movements or even directly call API to buy/sell instruments. At the end we will discuss and later build a stream processing pipeline in the IDE using the ML model.

A practical guide to LLM-based AI agents

Room D6

Philipp Wendland, Deloitte Consulting

This hands-on workshop is designed to provide participants with an in-depth practical understanding of how to leverage Large Language Models (LLMs) to create intelligent AI agents. As the rise of generative AI continues to transform industries it’s becoming increasingly important for both AI professionals and business leaders to understand the capabilities and implementation strategies for LLM-based systems. The Deloitte AI Institute focusses on brining AI expertise to clients across all industries ranging from innovation over strategy to capability building and scaling. Phillips’s strong technical background in physics / computer science enables him to bridge the gap between business and technology. 1: Introduction to LLM-based agents - Overview of the concept and architecture of AI agents - Introduction to popular frameworks to expedite agent development 2: Hands-on implementation - Guide participants through building a simple LLM-based AI agent using a given framework - Allow for customisation to demonstrate the flexibility and effectiveness of AI agents 3: Industry-Application - Outlook on newest developments of AI agents across various industries - Outlook on the potential of generative Ai and AI agents in particular across industries By the end of the workshop participants will have a solid understanding of the concepts behind LLM-based AI agents and hands-on experience in building these systems themselves. Led by an experienced facilitator from Deloitte’s AI Institute this workshop promises to provide valuable skills and knowledge that participants can leverage in their careers. This workshop is ideal for AI practitioners developers and researchers eager to explore the latest advancements in generative AI and apply LLM-driven automation in their respective fields.

Synthetic Data Generation for Embedding Model Fine-Tuning

Room D7

Ondřej Finke, O2/Dataclair
Stefan Josef, O2/Dataclair
Filip Roskovec, O2/Dataclair

Retrieving information from documents in non-English and domain-specific languages presents a challenge for many organizations. While general embedding models are powerful they often fall short when dealing with specialized terminology not encountered in their training data. This workshop offers a practical approach to addressing these issues: using a combination of real and synthetic data to build robust datasets for fine-tuning open embedding models. The workshop consists of two parts. First we provide an overview of embedding models fine-tuning techniques and methods for generating synthetic data tailored to these approaches. In the second part participants will engage in a hands-on session to generate synthetic data for fine-tuning their own models.

Saturday,
Workshops

O2 Universum, Českomoravská 2345/17a, 190 00, Praha (and on-line)

Registration from 9:00

Program will be announced.

Sunday,
Conference day 1

O2 Universum, Českomoravská 2345/17a, 190 00, Praha (and on-line)

Doors open at 08:30

Program will be announced.

Have a great time Prague, the city that never sleeps

You can feel centuries of history at every corner in this unique capital. We'll invite you to get a taste of our best pivo (that’s beer in Czech) and then bring you back to the present day at our networking event.

Venue ML Prague 2025 will run hybrid, in person and online!

The main conference as well as the workshops will be held at O2 Universum.

We will also livestream the talks for all those participants who prefer to attend the conference online. Our platform will allow interaction with speakers and other participants too. Workshops require intensive interaction and won't be streamed.

Conference building

O2 Universum
Českomoravská 2345/17a, 190 00, Praha 9

Workshops

O2 Universum
Českomoravská 2345/17a, 190 00, Praha 9

Now or never Registration

Early Bird

Only 85 Registrations Left

Standard

500 Registrations Until March 31

  • Conference days € 290
  • Only workshops € 230
  • Conference + workshops € 490

Late

Until sold out

  • Conference days € 320
  • Only workshops € 260
  • Conference + workshops € 520

What You Get

  • Practical and advanced level talks led by top experts.
  • Networking and drinks with speakers and people from all around the world.
  • Delicious food and snacks throughout the conference.

They’re among us We are in The ML Revolution age

Machines can learn. Incredibly fast. Faster than you. They are getting smarter and smarter every single day, changing the world we’re living in, our business and our life. The artificial intelligence revolution is here. Come, learn and make this threat your biggest advantage.

Our Attendees What they say about ML Prague

Thank you to Our Partners

Strategic Partners

Platinum Partners

Communities and Further support

Would you like to present your brand to 1000+ Machine Learning enthusiasts? Send us an email at info@mlprague.com to find out how to become a ML Prague 2025 partner. Our basic partnership offer can be found here.

Become a partner

Happy to help Contact

If you have any questions about Machine Learning Prague, please e-mail us at
info@mlprague.com

Organizers

Jiří Materna
Scientific program & Co-Founder
jiri@mlprague.com

Teresa Pulda
Event production
teresa@mlprague.com

Gonzalo V. Fernández
Marketing and social media
gonzalo@mlprague.com

Jona Azizaj
Partnerships
jona@mlprague.com

Ivana Javná
Speaker support
ivana@mlprague.com

Barbora Toman Hanousková
Communication
barbora@mlprague.com

Jan Romportl
Moderator