Hotel Metropol, Belgrade

16th-20th November 2026

DSC Europe 25 SCHEDULE

Ivo Andric B

NLP & LLMs,
Computer Vision & Robotics

Registration

09:00

 –

09:30

Registration

Building AI-Centric Apps with ADK, Gemini & Vertex AI: from Database explorers to Travel assistants

NLP & LLMs
This talk will showcase how Google’s Agents Development Kit (ADK), combined with Gemini and Vertex AI, can power real-world AI-centric applications. Through two demos—a SQL Agentic solution for natural language interaction with databases, and an AI travel assistant that plans city travel—we’ll explore how agentic architectures unlock scalable, interactive, and practical AI solutions.
09:30

 –

10:00

Building AI-Centric Apps with ADK, Gemini & Vertex AI: from Database explorers to Travel assistants

Gabriel Preda
Principal Data Scientist @ Endava

From Pre-Trained to Purpose-Built: Fine-Tuning LLMs for High-Impact Generative AI and Agents

NLP & LLMs
Technical talk, Business talk
Intermediate to Advanced
Fine-tuning pre-trained large language models (LLMs) offers a powerful and cost-efficient pathway to create high-performance, domain-specific Generative AI solutions. This session provides you with the practical knowledge and methodologies needed to harness fine-tuned models effectively, focusing on the complete lifecycle from data preparation to deployment. Participants will gain insights into how targeted, high-quality datasets can transform general-purpose models into specialized agents that excel in unique business and technical contexts. By the end of the session, attendees will understand both the strategic value and hands-on techniques required to translate fine-tuning into production-ready results.
10:00

 –

10:30

From Pre-Trained to Purpose-Built: Fine-Tuning LLMs for High-Impact Generative AI and Agents

Harshvardhan Jain
EMEA Generative AI Lead, Office of the CTO @ Oracle

Inside the AI Factory: Solving Financial Fraud with NVIDIA

10:30

 –

11:00

Inside the AI Factory: Solving Financial Fraud with NVIDIA

Nikola Vurdelja
Business Development Manager for AI and Cloud Services @ Orion telekom

The Control Stack: Building Guardrails for Enterprise LLM Applications

NLP & LLMs
This talk introduces a four-layer framework for controlling LLM outputs in production: input sanitisation, activation steering, output validation, and feedback loops. I’ll demonstrate how techniques from AI safety research – particularly activation steering – can guide model behaviour without retraining, and provide practical examples for each layer. Attendees will leave with actionable strategies for building more reliable LLM systems.
11:00

 –

11:30

The Control Stack: Building Guardrails for Enterprise LLM Applications

Egor Krasheninnikov
Generative AI Scientist @ Amazon Web Services (AWS)

Coffee Break

11:30

 –

12:00

Coffee Break

KEYNOTE: How Global Tech Leaders Drive Change Locally

Keynote
What does it take to turn global innovation into local impact? In this discussion, leaders from AMD, Microsoft, and Rivian share how world-class technology, AI, and software development are shaping Serbia’s role in the global tech ecosystem. The panel explores how multinational companies cultivate talent, foster innovation, and build sustainable growth at the intersection of global vision and local expertise.
12:00

 –

12:45

KEYNOTE: How Global Tech Leaders Drive Change Locally

Panelists: Djordje Simic, Drazen Sumic, Erhun Arkan
Director of Software Development and AMD Serbia Country Manager @ AMD, Partner Director @ Microsoft Serbia, Europe General Manager & Vice President Consumer Applications @ Rivian and Volkswagen Group Technologies

KEYNOTE: When Insight Isn’t Enough

Keynote
Explores the frustrating disconnect between analytics teams delivering great insights and decision-makers who don’t act on them. Discusses causes like risk aversion, misaligned incentives, and communication gaps.
12:45

 –

13:30

KEYNOTE: When Insight Isn’t Enough

Panelists: Natasa Radjenovic Zivanovic, Miroslav Grbovic, Valentin Konja
Head of Data Management, AI and Innovation Directorate @ OTP Bank, Head of CRM @ Banca Intesa, Director – Business Operations @ NCR Atleos

Lunch Break

13:30

 –

14:30

Lunch Break

Applied Sports AI: Turning Padel Match Footage into Actionable Insights

Computer Vision & Robotics
Technical talk, Business talk
Intermediate
Padel is one of the fastest growing sports in the world, with 30M players worldwide and a 17% YoY growth. Padel is so popular because it’s easy to pick up and enjoy, yet offers a long and exciting learning curve. Growth is important to padel players and Padelytics is a companion on that journey – an AI coach you can take with you to every match. In this presentation, we’ll introduce the sport of padel, explain our motivations for embarking on the journey of helping players improve and then go into a high level product and technical demonstration of the solution we’ve built, giving the audience a sneak peek into our product, as well as how it works under the hood.
14:30

 –

15:00

Applied Sports AI: Turning Padel Match Footage into Actionable Insights

Nenad Zivic
CEO and Head of Product @ Padelytics

Real-World Applications of Computer Vision in Automotive Systems

Computer Vision & Robotics
Technical talk
Intermediate
Computer vision has become a cornerstone of innovation in the automotive industry, powering everything from driver-assist technologies to intelligent vehicle features. Combined with advances in machine learning, it enables cars to interpret their surroundings, make informed decisions, and deliver safer, smarter driving experiences. This talk will provide an overview of how computer vision is applied across the automotive domain today. The second part will focus on a real-world example: Rivian’s Gear Guard feature. This feature illustrates how end-to-end machine learning and computer vision pipelines are developed, iterated, and scaled for deployment in production vehicles. By walking through its evolution—from initial implementation to added functionalities and improvements—the talk will highlight the challenges and cycles involved in bringing a computer vision feature from concept to a robust, user-facing system.
15:00

 –

15:30

Real-World Applications of Computer Vision in Automotive Systems

Jovan Sumarac
Computer Vision Engineer @ RV Tech

Coffee Break

15:30

 –

16:00

Coffee Break

Deep Learning for Mammography

Computer Vision & Robotics
Research talk
Intermediate to Advanced
Breast cancer is the most common type of cancer in women and has caused 670000 deaths globally in 2022. Serbia has the largest number of deaths caused by this type of disease in Europe, largely due to late diagnosis. Mammography is the gold standard for the detection and diagnosis of breast cancer. The procedure can be significantly enhanced with Artificial Intelligence (AI)-based software, which assists radiologists in identifying abnormalities. In this talk I will discuss several avenues that our research group is pursuing to push the state-of-the-art when it comes to application of computer-vision approaches in mammography.
16:00

 –

16:30

Deep Learning for Mammography

Dubravko Culibrk
Full Professor @ Faculty of Technical Sciences, University of Novi Sad

There Is No Spoon: Inferring Vision from Neural Codes. Reconstructing images from the brain’s internal model of reality.

Computer Vision
Technical talk, Research talk
Intermediate
A deep learning approach for reconstructing visual images from brain activity, developed as part of a master’s research project. By decoding neural representations into a latent visual space, the work explores how the brain encodes perception and how AI may approximate what the mind sees, with potential applications in brain–computer interfaces.
16:30

 –

17:00

There Is No Spoon: Inferring Vision from Neural Codes. Reconstructing images from the brain’s internal model of reality.

Dusan Pavlov
MSc Candidate, Neurocomputational Deep Learning Researcher @ Faculty of Sciences, University of Novi Sad, Independent Researcher

Official Closing

17:00

 –

17:15

Official Closing

Ivo Andric A

AI Infrastructure,

AI Software

Registration

09:00

 –

09:30

Registration

Democratizing AI with Affordable HPC Infrastructure

AI Infrastructure
Technical talk, Research talk
Beginner to Intermediate
When it comes to the AI infrastructure we are currently in a repeat of the 1970s era of computing. Compute resources are a hot commodity, typically offered by centralized data centers and paid for by the second. For commercial production use, this is not too bad, since everyone uses cloud anyway. However, education, for example, requires hands-on experience, iteration, failing fast and often, and it is difficult to achieve that when you must pay by the second. Affordable computing infrastructure, such as consumer GPUs, general purpose SoCs, AI-enabled CPUs, and “shoebox supercomputers”, such as Nvidia DGX Spark, offer a way to democratize AI through lowering the financial entry bar to the AI game. This talk will therefore focus on affordability in AI computing and its importance for educating the next generation of HPC and AI engineers.
09:30

 –

10:00

Democratizing AI with Affordable HPC Infrastructure

Dusan Gajic
Associate Professor @ Faculty of Technical Sciences, University of Novi Sad & Co-Founder @ 42computing

AI – how to start small and grow in the future

AI Infrastructure
Discover how to start your AI journey with Red Hat — from first experiments to enterprise-scale deployment. We’ll explore how open-source innovation, trusted platforms like Red Hat OpenShift AI, and a flexible approach to infrastructure enable organizations to begin small, demonstrate value quickly, and scale AI securely for the future.
10:00

 –

10:30

AI – how to start small and grow in the future

Andrzej Kowalczyk
Principal Solution Architect @ Red Hat Poland

Beyond the Cloud: The Rise of Decentralized AI

AI Infrastructure
TBA
10:30

 –

11:00

Beyond the Cloud: The Rise of Decentralized AI

Milovan Medojevic
Intelligent Production Systems Research Group Lead @ The Institute for Artificial Intelligence of Serbia

Unlocking AI potential with NetApp and your data

AI Infrastructure
11:00

 –

11:30

Unlocking AI potential with NetApp and your data

Jorge Gomez Navarrete
Technical Solution Architect @ NetApp

Coffee Break

11:30

 –

12:00

Coffee Break

Panel: Data, Gender and Justice in the Age of AI

Panel
Data, Gender and Justice in the Age of AI” explores how artificial intelligence and data-driven technologies shape gender dynamics and social justice. The panel examines issues such as algorithmic bias, representation in datasets, and the impact of AI on equity and inclusion, while highlighting strategies for building fairer and more accountable technologies.
12:00

 –

12:45

Panel: Data, Gender and Justice in the Age of AI

Moderator: Bojana Todorovic, Panelists: Petar Colovic, Biljana Stefanovic, Natasha Savic
Association @ AWDS, Associate Professor @ University of Novi Sad, Faculty of Philosophy, Department of Psychology Enterprise Architect – AI Strategy @ Atos, Staff AI Engineer @ Databricks

Black

12:45

 –

13:30

Black

Lunch Break

13:30

 –

14:30

Lunch Break

Agentic AI in Production: Lessons from Real-World Deployments

AI Software
Deploying agentic AI applications that actually deliver business value is hard, most teams are still figuring it out. In this session, Natasha will share how her team tackles these challenges head-on, taking AI agents from prototypes to production at scale. Through real-world case studies, you’ll gain insights, best practices, and lessons learned from building secure, scalable, and impactful agentic systems. Walk away with practical strategies to accelerate your own journey and avoid the pitfalls holding many teams back.
14:30

 –

15:00

Agentic AI in Production: Lessons from Real-World Deployments

Natasha Savic
Staff AI Engineer @ Databricks

Boy Scout Rule in the Age of AI: Leaving Codebase Cleaner than you Found it.

AI Software
Technical talk
Intermediate to Advanced
Artificial Intelligence is rapidly becoming embedded in developers’ workflows (from code-generation assistants to auto-refactoring bots) — and often the temptation is to rely on it blindly which can lead to messy and brittle code with a lot of duplication. This talk explores how “always leave the code cleaner than you found it” rule can be applied to modern code bases powered by AI. We will talk about concrete techniques like setting up and enforcing codebase standards, providing exhaustive context, linting, formatting, unit- and e2e-testing, refactoring code incrementally and keeping docs up to date. By the end, you’ll have a toolkit of strategies to make sure your AI-assisted code is not only working — but also remains maintainable and readable.
15:00

 –

15:30

Boy Scout Rule in the Age of AI: Leaving Codebase Cleaner than you Found it.

Georgii Perepechko
Staff Engineer @ Signal Ocean

Coffee Break

15:30

 –

16:00

Coffee Break

Beyond the Hype: Making AI Coding Assistants Actually Deliver Value

AI Software
AI coding assistants promise massive productivity gains, but most teams struggle to extract real value beyond the hype. The gap between promise and reality comes down to understanding how to work effectively with these tools—knowing when they amplify your capabilities versus when they create an illusion of progress. This talk shares practical insights on managing AI’s unpredictability, leveraging your expertise to guide and validate outputs, and maintaining control over production code quality. You’ll walk away with a clearer picture of what actually works, common pitfalls to avoid, and measured results from years of hands-on experience across the full evolution of AI coding tools.
16:00

 –

16:30

Beyond the Hype: Making AI Coding Assistants Actually Deliver Value

Marcos Heidemann
Principal ML/DS Engineer @ symphony.is

Building Automated Testing Pipelines for Multimodal LLMs

AI Software
Technical talk
Intermediate
How to develop LLMs testing framework that will guarantee that your system is healthy and with no regressions?
16:30

 –

17:00

Building Automated Testing Pipelines for Multimodal LLMs

Danijel Misulic
Senior Machine Learning Engineer @ Archetype AI

TESLA B – DSC:X

Building Data & AI Teams,

Data & AI Product Development

Registration

09:00

 –

09:30

Registration

Building the Future: The Role of Data Science Teams in AI Agent Development

Building Data & AI Teams
Business talk
Beginner to Intermediate
In the rapidly evolving world of AI, agents are at the forefront of driving innovation and automation. In this session, we’ll explore the different types of AI agents, from conversational bots to autonomous decision-making systems, and why data science teams are essential to their development. By diving into real-world use cases, we’ll demonstrate how cross-functional collaboration between data scientists, engineers, and domain experts plays a pivotal role in creating effective AI agents that meet business needs. We’ll also discuss the challenges and opportunities involved in adopting AI agents, including the skills and profiles required within teams to ensure seamless integration and scalability. Whether you’re new to AI or an experienced professional, this session will provide valuable insights into how data science teams are shaping the future of AI agents and driving their successful adoption across industries.
09:30

 –

10:00

Building the Future: The Role of Data Science Teams in AI Agent Development

Paula García Esteban
AI & Data Visualization Specialist @ StrataDom

Career building for data professionals

Building Data & AI Teams
Technical talk, Business talk
Intermediate to Advanced
People focus too much on technical skills for next paycheck, and not enough on planning and managing their career. Should we change that? We will demonstrate frameworks used for personal development. I hope you take and implement at least one idea from this lecture, and make your professional life better and more rewarding.
10:00

 –

10:30

Career building for data professionals

Josip Saban
Data Consultant

Behind the Curtain: How Data Roles Collaborate in the AI Era

Building Data & AI Teams
Business talk
Beginner
Data teams are often seen as a collection of separate roles (data engineers, analysts, data scientists, researchers, and AI engineers), each working in their own domain. But in practice, real impact comes from how these roles collaborate. In this talk, I’ll share how we at Nordeus build and structure our data teams to enable effective cross-functional work, with a focus on analysts and data scientists. I’ll cover the challenges we faced in aligning skill sets, expectations, and workflows, and how we developed ways of working that foster collaboration rather than silos. You’ll hear practical lessons on scaling analytics and ML capabilities, bridging the gap between research and production, and ensuring teams stay both innovative and business-focused.
10:30

 –

11:00

Behind the Curtain: How Data Roles Collaborate in the AI Era

Ekaterina Bubenko
Senior Business Analyst @ Nordeus

Coffee Break

11:00

 –

11:30

Coffee Break

KEYNOTE: Harnessing GEN AI in Fashion, Luxury and Beauty Specialized AI Solutions

Keynote
Use case, Solution showcase, Transformational talk
Intermediate to Advanced
In the competitive luxury and fashion industry, Artificial Intelligence has emerged as a strategic asset that enhances creativity, optimizes operations, and delivers highly personalized customer experiences. Learn how to leverage AI to drive your brand. Discover a curated selection of solutions that could enhance the brand and the teams.
11:30

 –

12:15

KEYNOTE: Harnessing GEN AI in Fashion, Luxury and Beauty Specialized AI Solutions

Raul Cruz Bonilla
Founder & CEO Brand transformation Leader @ Care Brand Management / Fermat / Kahoona

KEYNOTE: Industrial data as the fuel for the AI revolution in the oil and gas industry

Keynote
12:15

 –

13:00

KEYNOTE: Industrial data as the fuel for the AI revolution in the oil and gas industry

Nikola Nikacevic & Dalibor Lazarevic
Senior Account Manager @ Things Solver & Chief Data Officer @ NIS

Lunch Break

13:00

 –

14:30

Lunch Break

Black

13:30

 –

14:00

Black

Future-Proof Your Business: AI, Market Fit, and the Path to Predictable Profits

Panel
Business talk
Intermediate to Advanced
Capitalizing on AI and achieving the right product-market fit can significantly improve a business’s profit margin, predictability, and long-term sustainability.
14:30

 –

15:00

Future-Proof Your Business: AI, Market Fit, and the Path to Predictable Profits

Assem Hussein
Founder @ Valyr

Product Professional’s Journey to Full-Stack Product Developer

Data & AI Product Development
Transformational talk
Beginner
The role of a Product Manager is evolving. In a world where AI, automation, and no-code tools are reshaping the landscape, the next-generation PM isn’t just defining products—they’re building them. This talk explores the rise of the Full-Stack Product Developer, a hybrid professional who blends strategic thinking with hands-on execution. We’ll dive into how AI, APIs, and no-code platforms are empowering PMs to prototype, automate, and ship faster than ever. Whether you’re a PM looking to level up or a founder seeking an edge, this session will show you what’s next in product management. With a personal twist and hands on examples, exploring how (typically) non tech people can not only survive but thrive with DS, AI and powerful possibilities they unlock for everyone.
15:00

 –

15:30

Product Professional’s Journey to Full-Stack Product Developer

Milos Belcevic
Staff Product Manager @ Urban Sports Club

Coffee Break

15:30

 –

16:00

Coffee Break

AI Product Canvas: From Business Goals to Technical Decisions

16:00

 –

16:30

AI Product Canvas: From Business Goals to Technical Decisions

Mikhail Rozhkov
Technical Product Manager @ Nebius

How Product Data Scientists Can Use ML to Find Growth Opportunities

Most teams use ML to automate decisions; this talk shows how to use it to systematically discover growth opportunities. Daniil Demitrov (Head of Product Analytics, Uzum) presents a practical framework where a combination of ML methods acts as a hypothesis engine rather than an end product. You’ll learn the essential steps of implementing an ML-based hypothesis generator — starting with a research design document and building a dataset focused on actionable features (levers you can actually change), and ending with converting insights into prioritized experiments. The framework’s implementation reduces the time and cost of formulating pre-validated hypotheses by quickly ruling out weak candidates and focusing on changes that truly move business metrics.
16:30

 –

17:00

How Product Data Scientists Can Use ML to Find Growth Opportunities

Daniil Demitrov
Head of Product Analytics @ Uzum Fintech

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

* €10 from each ticket will be dedicated to a local humanitarian initiative, reflecting our commitment as an AI ecosystem to creating meaningful impact for those in need.