CO-LOCATED EVENTS

Hotel Metropol, Belgrade

17th-21st November 2025

DSC Europe 24 SCHEDULE

Ivo Andric B

Generative AI

Registration

08:30

 –

09:00

Registration

Official Opening

09:00

 –

09:15

Official Opening

Keynote: 7 Deadly Sins of AI (And How to Avoid Them)

Keynote
Business talk
Intermediate
AI is changing business as we know it – but are we making the same old mistakes with this new technology? In this talk, AI expert and author Tobias Zwingmann will reveal the seven most costly AI pitfalls that derail even the best companies. You’ll learn how to avoid these sins and implement AI in a way that’s profitable, not a cause for confession.
09:15

 –

10:00

Keynote: 7 Deadly Sins of AI (And How to Avoid Them)

Tobias Zwingmann
Author & Managing Partner @ RAPYD.AI

Marketing Hyper-personalisation with digital platforms and Generative AI

Generative AI
Technical talk, Business talk, Educational talk
Intermediate
In the dynamic marketing landscape, the rise of generative AI has opened up a multitude of possibilities. At the forefront of this technological advancement is hyper-personalisation, which allows brands to tailor their marketing efforts to the individual customer. By using generative AI to create hyper-personalised content at scale, companies can not only optimise their internal processes but also transform the customer experience. In this session, we provide business and technical insights with guidance on the use of generative AI in the field of hyper-personalisation in marketing and showcase our ground-breaking recent work in this field.

Marketing hyper-personalisation with GenAI is an automated solution to create highly personalized marketing promotional content (e-mail, mobile app push message) on a micro-segment level given a promotional goal and tone of voice, based on customer insights and characteristics from historical customer transactions and other customer activities.

The solution is comprising a micro-segmentation engine using AI based segmentation techniques, a Marketer Workbench UI for all marketer interaction activities custom built on Salesforce platform and an engine to create micro-segment based personas with marketing personalization content generation using Generative AI (watsonx) with direct integration to marketing campaign tool (Salesforce Marketing Cloud) to execute hyper-personalised marketing campaigns.
10:00

 –

10:30

Marketing Hyper-personalisation with digital platforms and Generative AI

Darije Ramljak
Associate Partner @ IBM Consulting

Fine-tune LLMs from Kaggle Models using (Q)LoRA

Generative AI
Technical talk
Beginner to Intermediate
The talk will introduce you on how to use Kaggle environment to perform fine-tunning of Large Language Models available as Kaggle Models using (quantized) Low-Rank adaptation method.
10:30

 –

11:00

Fine-tune LLMs from Kaggle Models using (Q)LoRA

Gabriel Preda
Principal Data Scientist @ Endava

Coffee Break

11:00

 –

11:30

Coffee Break

Boosting Data Efficiency with AI: Insights from a Case Study

Generative AI
Unlock the power of AI to transform customer service! Join Christian Casari from ONTEC AI at DSC Belgrade to explore how AI innovations revolutionized Sunny Cars’ customer support. Learn how leveraging cutting-edge solutions, such as email automation, knowledge management, and AI-driven insights, reduced response times, boosted efficiency, and enhanced customer satisfaction. This case study showcases actionable strategies to streamline business processes and embrace AI effectively. Don’t miss this opportunity to discover the tangible impact AI can have on modern business challenges!
11:30

 –

12:00

Boosting Data Efficiency with AI: Insights from a Case Study

Christian Casari
Head of AI Sales @ ONTEC

Beyond GPUs: How Intelligent Compute Accelerators (ICAs) are Shaping the Future of HPC and AI

Generative AI
Technical talk
Advanced
Modern AI technology is evolving from CNN through transformer networks to something currently unknown. NextSilicon architecture has no rigid processing units with imposed limitations. Contrary to that, it is dataflow-based and reconfigurable, even at runtime. Therefore, NextSilicon adapts to the required types of computing, from the most complex HPC applications to any future workloads.

NextSilicon can accelerate science — even the most challenging HPC applications — and will continue to be able to accelerate future workloads, including neural networks, given our flexible approach to acceleration. 2027 will see a continued evolution of network architecture, and NextSilicon is the best positioned in the market — with its patent-protected technology — to continuously adapt to this evolution.

Attendees will gain insights into how these changes shape the future of HPC and AI. Join us to discuss the broader implications of intelligent acceleration and its innovative adaptability to future workloads.
12:00

 –

12:30

Beyond GPUs: How Intelligent Compute Accelerators (ICAs) are Shaping the Future of HPC and AI

Aleksandar Beric
VP of R&D Serbia @ NextSilicon

Keynote: Understanding Text to Image Models

Keynote
Technical talk
Advanced
This presentation will dive into the fundamentals of text-to-image models, focusing on how systems like Stable Diffusion utilize CLIP-based embeddings to align text and image representations effectively. We’ll explore the role of CLIP in capturing nuanced relationships between modalities and examine sparse autoencoders as a tool for enhancing interpretability, offering insight into model internals and the potential for controlled image generation. Finally, we’ll discuss how sparse autoencoders enable steering of model outputs, paving the way for targeted, user-guided creation in generative art.
12:30

 –

13:15

Keynote: Understanding Text to Image Models

Jose Portilla
Instructor @ Udemy

Lunch Break

13:15

 –

14:30

Lunch Break

Keynote Fireside chat: LLMs – application and impact on business

Keynote Fireside chat
14:30

 –

15:15

Keynote Fireside chat: LLMs – application and impact on business

Sacha Khanin
CEO @ Polynome

How to train a large diffusion model

Generative AI
Technical talk
Advanced
Diffusion models represent the state-of-the-art approach to image generation. Models such as Midjourney, DALL-E 3 and Stable Diffusion are capable of generating high-quality images. They are also flexible and can be used for other related tasks such as inpainting, remixing and image animation. However, details on training these models are often sparse and incomplete.
The report outlines the typical requirements for training a large diffusion model and explains some common pitfalls.
15:15

 –

15:45

How to train a large diffusion model

Eugene Lyapustin
Senior ML Engineer @ Yandex

Leveraging multimodal language model for AI native wireless systems

Generative AI
15:45

 –

16:15

Leveraging multimodal language model for AI native wireless systems

Ali Mokh
Researcher Concept AI @ Ericsson

Coffee Break

16:15

 –

16:45

Coffee Break

Multi-Agent AI: Transforming Business Operations and Value Creation

Generative AI
Business talk
Intermediate
We are at a pivotal moment in AI development: models are becoming increasingly powerful while the costs of implementation continue to fall. This creates immense opportunities for businesses to unlock new levels of efficiency and drive innovation.

In this session, we will explore these advancements through two case studies: an AI-driven solution for a medical examinations provider that streamlines document validation and quality control, and an AI-powered system in the airline industry that autonomously manages social media interactions, classifying customer concerns, analyzing feedback, and enhancing service delivery.

The common thread in these examples is the transformative potential of autonomous AI agents—systems capable of intent analysis, strategic planning, and independent execution. These agents go beyond being tools; they enable sustainable automation and operational efficiency, positioning businesses to stay competitive in an ever-evolving landscape.

The time to act is now. The real question is not whether AI will replace roles, but how quickly your competitors will leverage it to outpace you. Partnering with the right experts is essential to navigating the AI ecosystem and securing your future in the business world.
16:45

 –

17:15

Multi-Agent AI: Transforming Business Operations and Value Creation

Radu Orghidan
Global SVP, Data & AI Strategy @ Endava

Exploring the Mechanisms of LLM Reasoning Capabilities

Generative AI
Technical talk, Business talk
Beginner to Intermediate
In this talk, we will explore the reasoning processes behind Large Language Models (LLMs) and examine whether their performance on various tasks reflects genuine human-like reasoning or if it is primarily driven by pattern-matching and memorization. By dissecting the mechanisms that enable LLMs to generate responses, we aim to provide a clearer understanding of how these models interpret, process, and generate information. This discussion will shed light on the current capabilities and limitations of LLMs, offering insights into their potential applications and the future of AI reasoning.
17:15

 –

17:45

Exploring the Mechanisms of LLM Reasoning Capabilities

Arsenii Shkunkov
Machine Learning Engineer @ ABBYY

Will an LLM replace me? A systematic study of Gen AI’s impact on data jobs

Generative AI
Business talk
Intermediate
As Generative AI, particularly Large Language Models (LLMs), continues to evolve, many in the data field are asking: Will an LLM replace me? In this talk, we’ll present a systematic study of data projects at Argusa to explore how LLMs are transforming tasks in data science, analytics, and engineering. Are these models a threat, an opportunity, or a bit of both? We’ll also highlight the skills that will become increasingly valuable in this new era. Join us to explore the future of data roles and their evolution with this cutting-edge technology.
17:45

 –

18:15

Will an LLM replace me? A systematic study of Gen AI’s impact on data jobs

Fatima Soomro
Head Of Operations @ Argusa

Day Recap

18:15

 –

18:30

Day Recap

Ivo Andric A

Data Engineering,

Decision Intelligence

Registration

08:30

 –

09:00

Registration

Black

09:00

 –

09:15

Black

Black

09:15

 –

10:00

Black

Supercharging Your Data Strategy: Building a Scalable Pipeline for AI & Web Insights with GitLab, Snowplow and Snowpipe

Data Engineering
Technical talk, Educational talk
Intermediate
In this talk, we will dive deep into GitLab’s journey to build a highly scalable and efficient data pipeline that captures both AI productivity metrics and GitLab web events. We will explore how we leverage Snowplow, an enterprise-grade event tracking platform, to collect and enrich raw event data, and Snowpipe infrastructure, Snowflake’s continuous data ingestion service, to seamlessly load this data into our Snowflake data warehouse.
10:00

 –

10:30

Supercharging Your Data Strategy: Building a Scalable Pipeline for AI & Web Insights with GitLab, Snowplow and Snowpipe

Ved Prakash
Staff Data Engineer @ Gitlab

Importance of Data Maturity before embracing AI

Data Engineering
Business talk, Educational talk
Intermediate
Understanding your company’s data maturity is essential before diving into AI implementation. This talk outlines key traits of mature data companies, such as data accessibility, integration, governance, and the importance of aligning people, processes, and technology to build a solid foundation for advanced analytics and AI.
10:30

 –

11:00

Importance of Data Maturity before embracing AI

Dejan Djekic
CEO @ Functor

Coffee Break

11:00

 –

11:30

Coffee Break

The State of Data Engineering

Data Engineering
Educational talk
Intermediate to Advanced
This talk will review the data industry’s history to see where we’re heading. This starts with data warehousing and goes into data science, and then, I finish off by showing how data engineering can avoid the same fate as data warehousing and data science.
11:30

 –

12:00

The State of Data Engineering

Jesse Anderson
Managing Director @ Big Data Institute

The Waiting Game: How CI/CD Disasters and Endless PR Delays Are Costing You Time, Money, and Sanity

Data Engineering
Technical talk, Educational talk
Intermediate
In today’s fast-paced business environment, Continuous Integration/Continuous Deployment (CI/CD) systems are meant to streamline development. Yet, too often, organizations find themselves trapped in endless pull-request (PR) delays and avoidable disasters. The hidden costs of inefficient pipelines will be explored—ranging from lost productivity to increased operational costs—while providing actionable insights to prevent these pitfalls. Drawing on real-world experiences from AWS PR Validation processes, we dive deep into how resolving these issues can save valuable time, reduce financial waste, and ease the frustration of your dev team. Join us to learn how to regain control of your CI/CD and boost your teams productivity!
12:00

 –

12:30

The Waiting Game: How CI/CD Disasters and Endless PR Delays Are Costing You Time, Money, and Sanity

Darija Strmecki & Lamia Cero
Senior Data Engineer @ Syntio & Data Engineer @ Syntio

Black

12:30

 –

13:15

Black

Lunch Break

13:15

 –

14:30

Lunch Break

Panel: Building Sustainable, Efficient, and Highly Intelligent AI Factories in Europe

Panel
As the world rapidly advances into the age of artificial intelligence, Europe’s position in the global AI landscape is more pivotal than ever. With ambitious plans to invest significantly in AI research and infrastructure—often likened to creating a “CERN for AI”;—Europe is poised to become a leader in developing responsible and cutting-edge AI technologies. This massive investment isn’t just about technological advancement; it’s about building a sustainable and efficient digital ecosystem that can serve not only the 500 million Europeans but also extend its benefits globally. Central to this vision is the development of innovative data centers and high-speed connectivity solutions that provide low-cost, high-bandwidth throughput. These are essential components for processing the vast amounts of data that AI systems require, all while maintaining a commitment to sustainability and reliability.
14:30

 –

15:15

Panel: Building Sustainable, Efficient, and Highly Intelligent AI Factories in Europe

Moderator: Tarry Singh, Panelists: Andrew van der Haar & Thomas Schmidt & Gauvain Girault
CEO @ Deepkapha & Managing Director @ Fiber Carrier Association & CEO and Co-Founder @ Amber AI & AI and Business Computing CTO – BDS @ Eviden

Data Driven Culture in Software Defined Vehicles

Decision Intelligence
Business Talk
Intermediate
This presentation will feature Rivian and its innovative use of data and AI to enhance vehicle development and software integration. Erhun’s presentation will explore how data-driven strategies are applied to improve vehicle software and connectivity, driving the future of electric vehicle technology.
15:15

 –

15:45

Data Driven Culture in Software Defined Vehicles

Erhun Arkan
General Manager of Serbia & Vice President of Rivian’s Global Consumer Application Software Teams @ Rivian

Alisa AI: Future-Proofing Your Business with Demand Forecasting

Decision Intelligence
Business Talk
Beginner to Intermediate
Unlock the power of predictive analytics with BE-terna Alisa AI solution. Our lecturer Nino Požar will present how advanced algorithms and Alisa AI can accurately predict future demand, optimize inventory, and drive profitability. Featuring a real-world case study, learn practical strategies to implement demand forecasting and enhance decision-making in business.
15:45

 –

16:15

Alisa AI: Future-Proofing Your Business with Demand Forecasting

Nino Pozar
Product Owner & Data Scientist @ BE-terna

Coffee Break

16:15

 –

16:45

Coffee Break

AI-Powered Insights Mining: CBS BrandPulse Solves AI Hallucinations

Decision Intelligence
Business Talk
Beginner to Intermediate
Very few can argue that LLM based AI has shown great potential in simulating human by connecting available knowledge written in human language.
However, when it comes to analyzing data and defining insights from it, without proper guidance LLM can often go in wrong directions and form wild paths, often leading to wrong conclusions.
We have integrated our data platform, CBS BrandPulse, based on rich data repository, with Open AI providing precise guidance to this LLM what data to read and how to interpret it.
Let us demonstrate how AI can form robust insights from the data (without wild rides), and even smartly hypotheses on Marketing and Sales data driven strategies.
16:45

 –

17:15

AI-Powered Insights Mining: CBS BrandPulse Solves AI Hallucinations

Milan Bjeloglav
CEO @ CBS Systems

Beyond the Hype: The Realities and Challenges of Implementing AI in Business

Decision Intelligence
Business Talk
Intermediate
While AI often enters the business arena as an exciting ‘Trojan Horse’ promising innovation and competitive edge, the real challenges lie hidden within. This talk will uncover how critical elements like data governance, data lakes, and robust data infrastructure are the true drivers behind successful AI integration. We’ll explore how to navigate these hidden complexities to move beyond the allure of AI and build a solid foundation for sustainable growth.
17:15

 –

17:45

Beyond the Hype: The Realities and Challenges of Implementing AI in Business

Martina Naumovska Necheska
Partner & Head of Data Science @ Data Masters

Business Decision-Making in the Era of AI

Decision Intelligence
Transformational talk
Beginner to Intermediate
This talk explores how AI transforms traditional business processes by streamlining data cleaning, preparation, and ETL tasks. This innovation diminishes the need for business analysts, enabling direct interaction with AI systems for reliable data insights. AI’s advanced capabilities allow it to strategize and provide meaningful, high-level advice, making data democratization a reality. Executives, managers, and all employees can now access and query data effortlessly, fostering a more inclusive and efficient decision-making environment within organizations.
17:45

 –

18:15

Business Decision-Making in the Era of AI

Benjamin Dzubur
AI Team Lead @ Zebra BI

Day Recap

18:15

 –

18:30

Day Recap

Lavender

Deep Learning,

ML Application

Registration

08:30

 –

09:00

Registration

Black

09:00

 –

10:00

Black

Beyond Supervised Learning with Zero-Shot Vision

Deep Learning
Technical talk
Advanced
Zero-shot image classification is evolutionizing computer vision by enabling models to classify images of unseen objects without prior training examples. This presentation explores the core principles of this technology, including its ability to leverage semantic knowledge and relationships between visual features and textual descriptions. We will delve into cutting-edge techniques like CLIP, discuss current limitations, and showcase real-world applications in fields like image retrieval and autonomous vehicles.
10:00

 –

10:30

Beyond Supervised Learning with Zero-Shot Vision

Nemanja Milosevic
Assistant Professor @ Department of Mathematics and Informatics, University of Novi Sad

The Double-Edged AI Sword: Examining Applications in Different Fields

Deep Learning
Research talk
Beginner
In this talk, we’ll explore various AI technologies, showcasing how they can be both valuable tools and potential risks. We’ll highlight examples from fields like transportation and computer vision to illustrate AI’s transformative power. While celebrating its potential, we’ll also address the need to regulate its usage. Join us as we discuss the dual nature of AI, emphasizing the importance of enabling its development while maintaining secure world.
10:30

 –

11:00

The Double-Edged AI Sword: Examining Applications in Different Fields

Sead Delalic
Assistant Professor @ University of Sarajevo & AI Consultant

Coffee Break

11:00

 –

11:30

Coffee Break

Hominis: Training Human-centric Large Language Models

Deep Learning
Technical talk
Advanced
This talk introduces Hominis, a groundbreaking family of large language models (LLMs) developed with a strong focus on human-centricity and environmental responsibility. We will explore the journey of scaling Hominis while maintaining a keen awareness of our environmental impact, detailing the innovative solutions implemented to address the challenges encountered during development. The presentation highlights the unique aspects of training Hominis on Leonardo, the European Union’s third-largest supercomputer, and discusses how this powerful infrastructure contributed to the model’s capabilities. The talk will conclude with a demonstration showcasing Hominis in action, allowing attendees to witness firsthand the practical applications and potential of this model.
11:30

 –

12:00

Hominis: Training Human-centric Large Language Models

Narendra Patwardhan
CTO @ DK AI Lab

Optimizations of AI inference: The role of ML compiler and Hardware innovations

Deep Learning
Technical talk
Advanced
In this presentation, we will explore the critical role of machine learning (ML) compilers and hardware innovations in optimizing AI inference. As AI models, particularly large language models (LLMs), continue to grow in complexity and size, the demand for efficient inference has never been higher. We will discuss the importance of ML compilers in translating high-level model descriptions into optimized hardware instructions, ensuring maximum performance and efficiency. Additionally, we will address the unique challenges posed by LLMs, such as high computational and memory requirements, and present strategies to overcome these hurdles through hardware advancements and compiler optimizations. Attendees will gain insights into the latest trends and technologies driving AI inference, including the use of specialized hardware like GPUs, NPUs, and custom accelerators, and how these innovations are shaping the future of AI deployment.
12:00

 –

12:30

Optimizations of AI inference: The role of ML compiler and Hardware innovations

Milan Stankic
Senior Engineering & Delivery Lead @ HTEC

Black

12:30

 –

13:15

Black

Lunch Break

13:15

 –

14:30

Lunch Break

Panel: AI development and copyright laws: public availability is not the same as free use

Panel
AI developers must navigate the complexities of copyright law when collecting and using data for training models. Why? Because when developing their own AI models, AI providers use various data sets. These data sets can be provided by clients, but the models may also be trained on licensed or publicly available data. However, public availability does not imply free use. Copyright compliance should be a primary consideration from the moment training data sets are being collected. Even if the data is publicly available (e.g., published on YouTube), it does not mean that such data can be freely used for training AI models. This is a common misconception among AI developers.

From the onset, developers should ensure they have the necessary permissions or legal basis to use the data. By adhering to copyright obligations, AI developers can mitigate legal risks and ensure compliance with relevant laws. Failure to comply with these laws, might not only trigger liability, but may also render the final model unmarketable in certain jurisdictions.

At DSC, we will delve into the world of copyright law as it pertains to data scraping in the age of AI, to explore the legal challenges and implications of using AI to collect and utilize data and focus on the balance between innovation and intellectual property rights. Startups and IT companies now heavily depend on data scraping for their operations, but many may not even be aware of the significant legal risks it triggers. This session is therefore essential for anyone looking to stay ahead of the curve in understanding the intersection of AI and copyright law.
14:30

 –

15:15

Panel: AI development and copyright laws: public availability is not the same as free use

Moderator: Ljubomir Kljajic, Panelists: Andrea Radonjanin & Branko Kovac & Dusko Martic
Patent Attorney @ Intellectual Property (IP) Systems & Partner in IP, Corporate, M&A, Regulatory @ Moravcevic, Vojnovic and Partners in cooperation with Schoenherr & AI Consultant @ Logikka & Head of Content Protection @ United Group

Building next-gen speech synthesis for Bookmate audiobook service

ML Application
Technical talk
Advanced
This presentation explores the development and evolution of text-to-speech (TTS) technology for audiobook production. We begin by detailing our baseline approach to creating a high-quality TTS system using only 90 hours of training data, which we successfully deployed across more than 100,000 books. We’ll share insights into voice selection, dataset construction, and meeting user expectations for synthetic speech.
The second part of our talk focuses on our leap to next-generation TTS, utilizing tens of thousands of hours of data. We’ll discuss the intricacies of large-scale dataset management, advancements in neural network architectures, and our methods for quality assessment through crowdsourcing and user feedback.
15:15

 –

15:45

Building next-gen speech synthesis for Bookmate audiobook service

Vladimir Platonov
TTS (text-to-speech) Department Lead @ Yandex

How AI Agents Will Change the Future of Programming

ML Application
Educational Talk
Intermediate
Nowadays, there are many speculations about the future of programming – some believe that programmers will soon be replaced by robots, while others argue that modern LLMs merely simulate thinking and are far from solving real problems. Meanwhile, the industry is rapidly moving towards improving the quality of programming tasks – new models (like o1), new startups tackling coding challenges, and swift advancements in agent-based benchmarks (including OS models). At Zencoder, we strive to stay at the forefront of progress and want to share our vision of AI in programming, AI Agents, and the future we foresee.
15:45

 –

16:15

How AI Agents Will Change the Future of Programming

Dmitry Krasnov
Engineering Manager @ Zencoder

Coffee Break

16:15

 –

16:45

Coffee Break

TIS Group’s AI Journey from Healthcare to Agriculture and then back to the roots

ML Application
Business Talk
Beginner to Intermediate
• Why does TIS.AI interact with infants? (sendd.eu)
• How do vineyards interact with TIS.AI? (vineyardangel.com)
• What is TIS.AI official LLM language?
• Where are our business roots?
16:45

 –

17:15

TIS Group’s AI Journey from Healthcare to Agriculture and then back to the roots

Tomislav Strgar
Head of AI Department @ TIS Group

Federated Data Architecture

ML Application
Federated Data Architecture is the concept which follows the main principles of new Data Mesh paradigm with small modifications considering the specificities and needs of UniCredit Bank Serbia. This allows us to make exact and clear business ownership together with Data Products development. Our goal is to fully support building of data driven company where everyone and everything has access to all the data required to make the best possible decisions. Data and AI will be used to learn at scale about our customers, our markets and ourselves and to adapt quickly to changes. This will allow us to deliver relevant and personalized experiences and advise with unparalleled time-to-market driven.
17:15

 –

17:45

Federated Data Architecture

Miodrag Pesovic
Head of Data and Intelligence @ UniCredit Bank

Our journey to harden the security and compliance of the Data Platform

ML Application
Technical talk
Intermediate to Advanced
Talk about our journey to harden the security and compliance of the Data Platform with the steps and learnings we have to date.
17:45

 –

18:15

Our journey to harden the security and compliance of the Data Platform

Dennis van Rooijen
Director Data Platform @ GitLab

Day Recap

18:15

 –

18:30

Day Recap

Tesla B – DSC:X

Data in Cloud,

AI & Data Strategy

Registration

08:30

 –

09:00

Registration

Black

09:00

 –

09:15

Black

Black

09:15

 –

10:00

Black

Scalable and Sustainable AI Infrastructure (On The Seabed)

Data in Cloud
Transformational talk
Intermediate to Advanced
As AI continues to revolutionize industries, the focus often remains on groundbreaking algorithms and applications. However, with Europe passing 150 TWh to power its data centres and 17 million households CO2 emission equivalent only for cooling – power and sustainability will be the limiting factor for our AI journey.
In this session, Peter Stensgård Hansen, former Ørsted and Senior Partner in Subsea Cloud Partners, will present how European offshore technologies is taking another step change toward energy scalability and sustainability by placing offshore data centres on the seabed. The technology to build offshore energy systems was started in the 1970s with oil and gas, during 2000s the first large scale offshore wind farms took over. Today, Europe can again be pioneers to leverage 50 years of offshore infrastructure experience to build co-located large scale offshore cost and energy efficient HPC data centres on the seabed.
Ideal for company executives and AI enthusiasts, this session underscores the pivotal role of infrastructure in the AI landscape and offers actionable insights to safeguard and enhance AI’s future.
10:00

 –

10:30

Scalable and Sustainable AI Infrastructure (On The Seabed)

Peter Stensgaard Hansen
Managing Partner @ Subsea Cloud Partners

Future of work with google workspace, be ready to start the journey now

Data in Cloud
Your Powerful AI assistant is here, Gemini powered by generative aI is transforming digital assistance by tackling complex tasks and saving you hours each day. In this presentation, Google’s Suzana Drakulic will show and discuss the potential of and future of Generative AI and explore the ways in which it can be used to enhance and even transform your way of working with focus on productivity and security.
10:30

 –

11:00

Future of work with google workspace, be ready to start the journey now

Suzana Drakulic & Vanja Melnik
Head of Google Workspace Nordic & Central Eastern Europe @ Google & Business Development Google Cloud Adriatics @ Devoteam Serbia

Coffee Break

11:00

 –

11:30

Coffee Break

Exploring LLMs across clouds – A Year in the Life

Data in Cloud
Transformational talk
Intermediate
Large Language Models (LLMs) are powerful tools for natural language processing and generation, and they have been widely adopted by various cloud service providers. However, the practical implications of the differences in the approaches, features, and limitations of LLMs across different cloud platforms are significant. In this session, we will compare and contrast the LLM offerings of Google, Microsoft, and Amazon and provide an overview of the pros and cons of each of them. We will also explore the different options and challenges of integrating LLMs in custom applications, as well as the tooling ecosystem that each vendor provides to help developers build LLM-based solutions. Finally, we will discuss the security and safety aspects of using LLMs in the cloud, and the best practices for ensuring ethical and responsible use of generative AI.
11:30

 –

12:00

Exploring LLMs across clouds – A Year in the Life

Tomislav Tipuric
CTO @ Nephos

Building an AI Platform

Data in Cloud
Technical talk, Business talk, Educational talk
Beginner
A story of a bunch of crazy people dreaming big, failing a lot, and constantly pivoting. Through it all we’ve learned a thing or two, and managed to create a platform that allowed us to launch dozens of AI products (even before GenAI was a thing). Finally, the platform matured enough to become a product itself, allowing us to scale our dreams globally.
12:00

 –

12:30

Building an AI Platform

Livio Bencik
Head of AI @ Mindsmiths

Keynote: How to successfully scale GenAI in big organizations

Learn how Deloitte supported their clients in going from GenAI proof of concepts to @scale solutions that generate business impact across the organization. You will learn what are the typical blind spots by data and tech teams in building the new solutions with GenAI. We will discuss key elements required for successful scale-up, see live demos of Deloitte’s GenAI solutions, and be able to ask questions to two Deloitte Partners focused on AI & Data capabilities: Amadeusz Andrzejewski & Jan Hejtmanek.
12:30

 –

13:15

Keynote: How to successfully scale GenAI in big organizations

Amadeusz Andrzejewski & Jan Hejtmanek
Partner & AI and Data Leader for Poland @ Deloitte & Partner @ Deloitte

Lunch Break

13:15

 –

14:30

Lunch Break

Keynote: Leveraging Data to Unlock Business Value

Financial Institutions and Merchants are leveraging data to drive desirable business outcomes by connecting deeply with customers. Visa helps its clients to unlock the full potential of data. In this session we will share few examples of how many global companies are addressing this challenge and how Visa uses data to effectively attract and engage clients, as well as to manage risk.
14:30

 –

15:15

Keynote: Leveraging Data to Unlock Business Value

Roberto De Meo
Vice President @ Visa Consulting & Analytics, Central and Eastern Europe, Middle East and Africa

Building Trust in AI: Alstom’s Journey Towards Safe, Reliable Mobility Solutions

AI & Data Strategy
In the mobility industry, the majority of AI efforts have focused on non-safety-critical applications—areas that, while important, represent only a fraction of the potential impact on our business. But imagine the transformative power of AI in safety-critical applications: autonomous trains, advanced signaling systems, and robust Automatic Train Protection and Operation (ATP and ATO).

However, the future of AI is not solely about innovation; it’s about trust. In this talk, we will explore how Alstom is building trust in AI for non-safety-critical applications today and discuss the road ahead toward safely integrating AI into the safety-critical systems of tomorrow.
15:15

 –

15:45

Building Trust in AI: Alstom’s Journey Towards Safe, Reliable Mobility Solutions

Nenad Mijatovic
Director of Data Science and AI @ Alstom

Fifty shades of Data & Analytics Strategy and 5 of Its Most Underrated Components

AI & Data Strategy
In today’s data-driven world, crafting a business-aligned data/AI strategy is essential yet complex. This session guides data managers through the key steps and critical challenges: aligning data initiatives with business goals to prevent overengineering, prioritizing impactful Data/AI/BI use cases, and fostering data/AI governance that evolves with organizational maturity. Attendees will gain actionable practices for resource management, designing an adaptive operational model. and effective tool adoption, establishing a foundation for sustainable data and analytics growth.
15:45

 –

16:15

Fifty shades of Data & Analytics Strategy and 5 of Its Most Underrated Components

Alex Barakov
Data & Analytics Manager @ Luxoft

Coffee Break

16:15

 –

16:45

Coffee Break

Real-World Strategy to Create Business Impact by Maximizing Data Usability

AI & Data Strategy
As Thomas Edison said in his talk, “The value of an idea lies in the use of it.” The same is true with data. Or, provocatively, your data is worth nothing unless you use it. How much do you use your data? How fast do you produce information? How fast does the information flow in your company? We will share practical solutions to align the people, processes, and solutions to the challenges that most companies face in their data journeys and share the steps to define and implement a data strategy.
16:45

 –

17:15

Real-World Strategy to Create Business Impact by Maximizing Data Usability

Filip Petrovski
Co-founder & CEO @ Data Masters

Why AI Projects Fail: The Missing Link of a Strong Data Strategy

AI & Data Strategy
Artificial Intelligence is often promoted as a game-changer, promising productivity gains, improved decision-making, and a competitive edge. Yet, many AI projects fail to deliver on this potential—not because of flaws in the technology, but due to a lack of alignment with a clear data strategy that supports business objectives. Without this strategic foundation, AI initiatives risk becoming costly experiments that fall short of driving real value.

In our session we’ll explore how organizations can create a strong data strategy to harness AI’s full potential. This ensures AI fits smoothly into workflows and solves real business challenges. It’s time to move beyond the hype and focus on building a data-driven culture that delivers measurable, strategic results.
17:15

 –

17:45

Why AI Projects Fail: The Missing Link of a Strong Data Strategy

Barbara Peruskovic & Linda Stege
Data Strategy Advisors @ Data Coase

Quadrupling Data Science Impact: removing bottlenecks from your data science process

AI & Data Strategy
Transformational talk
Beginner to Intermediate
Even as data science continues to mature within organizations, delivering tangible impact remains a common challenge. In this session, we’ll explore a case study from Bank of Georgia, where our team successfully overcame this challenge and quadrupled data science impact in a year into 8-figure territory. We’ll share our journey of transformation, discussing how we identified key bottlenecks and the structural changes, process improvements, and employee engagement strategies that were applied to mitigate them. Throughout the presentation, we’ll dive into specific examples and practical insights that you can apply to your own data science projects. We’ll cover topics such as optimizing workflows, fostering collaboration, and maintaining team morale and motivation. Whether you’re a data scientist, manager, or executive, we hope this session will provide takeaways to help you maximize the impact of data science in your organization.
17:45

 –

18:15

Quadrupling Data Science Impact: removing bottlenecks from your data science process

Mikheil Nadareishvili
Chief Analytics Officer @ Bank of Georgia

Day Recap

18:15

 –

18:30

Day Recap

* this program is not final and is subject to change, full schedule will be available soon