Hotel Metropol, Belgrade
18th-22nd November 2024
DSC Europe 24 SCHEDULE
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

Elytis
Intelligent Inventory Management,
Data-driven Product Recommendations,
AI in Marketing Automation
Data Driven Supply Chain Network Design
Intelligent Inventory Management
The idea of this topic is to illustrate how data is leveraged to support decision-making in one of the most critical areas of supply chain planning. Supply chain network design consists of various decision processes related to network optimization, specifically in determining the optimal number, locations, and sizes of distribution warehouses. Additionally, it defines how material flows are coordinated between facilities. This process encompasses various areas of analytics and data handling, including data engineering and report generation for descriptive analysis, utilizing cluster analysis to identify centers of gravity, and finally developing optimization models aimed at minimizing overall supply chain costs.
10:00
–
10:30
Data Driven Supply Chain Network Design
Mila Stankovic
Lead IT Technical Consultant @ Bosch
AI in Retail – Advancing Promo Demand Forecasting
Intelligent Inventory Management
Accurately predicting customer demand and efficiently scheduling supply chain operations are today’s critical differentiators in the retail landscape. This presentation will demonstrate how AI-driven demand forecasting and supply chain scheduling can not only reshape retail operations but also deliver measurable improvements in operational efficiency, long-term profitability, and sustainable competitive advantage.
Unlocking Precision in Marketing and Distribution:
Backed by extensive experience with major retail brands, where we achieved a 2.5-fold forecasting error decrease, this AI-driven demand forecasting solution aligns marketing efforts directly with consumer behavior. It ensures that promotions and product launches are perfectly timed and inventory is optimally distributed across store networks. This approach reduces the risk of overpromising and underdelivering, preserving brand integrity and maximizing the impact of every marketing euro.
Strategic Supply Chain Orchestration:
We’ll also explore how sophisticated supply chain scheduling solution ensures that deliveries from central warehouses to a vast network of stores are executed seamlessly, even during peak periods. By integrating demand forecasting with supply chain scheduling, retailers can better anticipate market fluctuations and allocate resources more effectively, resulting in enhanced agility and reduced operational strain.
Real-World Insights from Leading Retailers:
Based on our experience with large retail clients, in this talk, we will describe how integrating these AI-driven models transforms retail operations into becoming more anticipatory and efficient. By anticipating customer needs and aligning resources in advance, companies can exceed consumer expectations and sustain growth in a competitive market.
AI-driven demand forecasting and supply chain scheduling are crucial in enhancing retail operations, turning challenges into opportunities. This approach focuses on building resilience and ensuring that every decision, from marketing to distribution, contributes to long-term success and reinforces market leadership.
Unlocking Precision in Marketing and Distribution:
Backed by extensive experience with major retail brands, where we achieved a 2.5-fold forecasting error decrease, this AI-driven demand forecasting solution aligns marketing efforts directly with consumer behavior. It ensures that promotions and product launches are perfectly timed and inventory is optimally distributed across store networks. This approach reduces the risk of overpromising and underdelivering, preserving brand integrity and maximizing the impact of every marketing euro.
Strategic Supply Chain Orchestration:
We’ll also explore how sophisticated supply chain scheduling solution ensures that deliveries from central warehouses to a vast network of stores are executed seamlessly, even during peak periods. By integrating demand forecasting with supply chain scheduling, retailers can better anticipate market fluctuations and allocate resources more effectively, resulting in enhanced agility and reduced operational strain.
Real-World Insights from Leading Retailers:
Based on our experience with large retail clients, in this talk, we will describe how integrating these AI-driven models transforms retail operations into becoming more anticipatory and efficient. By anticipating customer needs and aligning resources in advance, companies can exceed consumer expectations and sustain growth in a competitive market.
AI-driven demand forecasting and supply chain scheduling are crucial in enhancing retail operations, turning challenges into opportunities. This approach focuses on building resilience and ensuring that every decision, from marketing to distribution, contributes to long-term success and reinforces market leadership.
10:30
–
11:00
AI in Retail – Advancing Promo Demand Forecasting
Bruno Segvic
Co-Founder & CEO @ IT FROM BIT
AI-Powered Pricing: Smarter Decisions, Bigger Profits
Intelligent Inventory Management
Use case
Beginner to Intermediate
In today’s competitive retail landscape, success depends on mastering precise pricing strategies—and AI is leading the way in transforming this process. This talk will explore how AI-powered pricing is surpassing traditional methods by analyzing vast amounts of data, identifying meaningful patterns, and making real-time adjustments that go far beyond human capability. From understanding price elasticity to predicting consumer behavior, AI eliminates the guesswork in pricing, enabling businesses to drive profitability and maintain a competitive edge. Learn why relying on intuition is no longer enough and how AI can empower companies to make smarter, data-driven pricing decisions that maximize revenue and performance.
11:30
–
12:00
AI-Powered Pricing: Smarter Decisions, Bigger Profits
Anthony Gale
Co-founder & CEO @ PromoLens
AI search solutions for online marketplaces: new experiences and technical aspects.
Data-driven Product Recommendations
Search solutions are changing across many business domains, enhanced by new AI-based functionality. At Velebit AI, we’ve been developing multiple computer vision and NLP solutions for online marketplaces for many years. This talk will share insights on how AI will change the e-commerce experience. We’ll go over multiple new multimodal AI search use cases for online marketplaces, along with the main trends and technical aspects.
12:00
–
12:30
AI search solutions for online marketplaces: new experiences and technical aspects.
Mladen Fernezir
Co-founder and Lead Data Scientist @ Velebit.ai
How to scale B2B sales operations? – Building AI systems that sell better
Data-driven Product Recommendations
Use case
Intermediate
The talk will explore the development of internal AI tools that automate outreach to merchants across multiple sales channels. It discusses a strategic approach to prioritizing accounts and how AI-generated content, combined with predictive machine learning models, has significantly boosted sales without the need for additional headcount. By sharing real-world examples and metrics, it aims to illustrate the tangible impact of these innovations on sales processes. Attendees will gain valuable insights into leveraging AI to enhance sales operations and drive efficiency in B2B environments, ultimately leading to improved business outcomes and a competitive edge in the market.
12:30
–
13:00
How to scale B2B sales operations? – Building AI systems that sell better
Agostino Calamia
Sr. Machine Learning Engineer @ Freelance
13:00
–
13:45
Panel: AI Solutions for Enhancing Customer Loyalty and Engagement
Moderator: Slavoljub Atanasovski, Panelists: Irena Bojarovska & Dusan Popovic & Milos Babic
Digital Product and Partnerships Manager @ HTEC & Applied Scientist @ Zalando SE & Co-founder and CEO @ Byteout & Head of Data @ Ananas
Machine Learning: The Engine Behind Modern E-commerce Search
Data-driven Product Recommendations
Technical talk, Research talk
Beginner to Intermediate
E-commerce platforms have become an integral part of our daily lives, yet we often overlook the complex processes occurring behind the scenes. From keyword matching to semantic search, re-ranking, recommendation systems, and personalization, these platforms employ a variety of sophisticated models, both predictive and generative. This talk will explore the key components of E-commerce search engines, breaking down their essential functions and underlying technologies. We’ll examine how these systems work together to deliver relevant results and enhance user experience, providing attendees with a clearer understanding of the technology powering modern online shopping.
15:15
–
15:45
Machine Learning: The Engine Behind Modern E-commerce Search
Milutin Studen
Machine Learning Research Engineer @ SmartCat
Demand Time-Series Forecasting: From Accuracy to Stability and Automation
AI in Marketing Automation
Use case
Intermediate
Founded in 2008, Zalando is a leading pan-European fashion and lifestyle e-commerce platform, serving about 50 million active customers across 25 markets. In this talk, I will take you through the journey of creating machine-based demand forecasts at a market level. We will explore what makes a “good” forecast by defining and measuring accuracy, and discuss the importance of stability, especially in real-world applications where consistency can be just as critical as accuracy. Lastly, I will explain how we deploy our models at scale using AWS SageMaker, Databricks, and Airflow, and how these forecasts drive decision-making in practice.
16:00
–
16:30
Demand Time-Series Forecasting: From Accuracy to Stability and Automation
Irena Bojarovska
Applied Scientist @ Zalando SE
Transforming E-Commerce with LLMs: 3 Real-World Use Cases
AI in Marketing Automation
Use case
Beginner to Intermediate
Join us as we explore how LLM-powered systems can revolutionize the automation of tedious and repetitive tasks, focusing on the Swiss e-commerce sector. We’ll discuss three innovative LLM-powered automation use cases Boxalino developed and implemented for its partners: automating SEO writing and marketing content generation for PerfectHair.ch, extracting product information from thousands of documents to redesign a product information management system for Lehner Versand, and connecting people in need of home renovation with the appropriate craftsmen through a project description writing assistant developed for renovero.ch. We’ll delve into the challenges faced during the development and implementation of these systems, as well as the substantial value they bring to users.
16:30
–
17:00
Transforming E-Commerce with LLMs: 3 Real-World Use Cases
Ratko Nikolic
Senior Data Scientist @ InterVenture
Optimizing E-commerce with Data Science: Enhancing Customer Experience and Sales
AI in Marketing Automation
Use case, Panel with the solution provider and the client, Transformational talk
Intermediate
This talk explores how data science transforms e-commerce by optimizing customer experiences, increasing sales, and streamlining operations. We’ll dive into key strategies like RFM and cohort analysis, recommendation systems, and A/B testing to improve conversion rates. Learn how leveraging these techniques not only boosts customer engagement but also drives sustainable growth. Ideal for e-commerce professionals looking to harness data to personalize their offerings and elevate their business performance.
17:00
–
17:30
Optimizing E-commerce with Data Science: Enhancing Customer Experience and Sales
Youssef Kamal
Founder @ People of Data – Middle East

Nikola Tesla A
Neuro AI
Modern Challenges Of Humanities For The Development Of AI: Neuromorphic Linguistics As A Reflection Of New Approaches To Understanding Emotional Intelligence For Solving Problems Of Unconventional And Neuro AI
The Fourth Industrial Revolution and the active use of various IT technologies have highlighted the importance of neuro-psychophysiological studies of language and speech, which allowed us to single out neuromorphic linguistics separately, making a significant contribution to understanding the dialogue between a person and his body, the emotional and emotive standardization of ethno-sociocultural ways of objectifying the world, as well as to the methodology of non-traditional engineering of emotivity of the speech and behavioral profile of natural and artificial intelligence, which allows you to focus on interdisciplinary approaches, combining neuroscience, NLP and AI. The main problem is caused by a non-discrete approach to the assessment of emotivity in communication and emotions in the behavioral matrix. The report is a development of the basic logic of our hypothesis and the work done on the gradation of emotional intelligence and, accordingly, emotional artificial intelligence with the following structure: general emotional intelligence, emotionogenic intelligence, neuromorphic intelligence and emotive intelligence. The modern appeal to the capabilities of the deep neurocognitive constructor in social engineering is based on the neuro-psycholinguistic mechanisms of human brain activity, which allows for a clearer understanding of the structure of general emotional intelligence. In this regard, a distinction is made between the concepts of ’emotion’, ’emotivity’, ’emotionality’, which is seen as important in creating a full-fledged emotional artificial intelligence. Neuromorphic linguistics is a deep analyzer of the language of communication based on a new interpretation of the biomimetic concept using various parameters of emotive-emotional modeling of neurosimulation, capable of teaching intelligent systems of emotivity with the subsequent creation of emotional-emotive artificial intelligence systems similar to human ones.
10:00
–
10:30
Modern Challenges Of Humanities For The Development Of AI: Neuromorphic Linguistics As A Reflection Of New Approaches To Understanding Emotional Intelligence For Solving Problems Of Unconventional And Neuro AI
Irina Karabulatova
Professor @ Lomonosov Moscow State University
Memristive Devices In Neuromorphic Systems
Memristive devices are considered as promising elements for the realization of neuromorphic systems, as they can mimic essential properties of biological synapses and can be used as parts of artificial neurons. Initially, we will consider briefly important features related to memristor architecture, materials, technology of their production and properties. Next part will be dedicated to the definition of neuromorphic systems and the comparison of some properties, playing positive role in neuromorphic systems but very negative in traditional electronic circuits (noise, cross-talk of elements, etc). In the next part we will make a comparison of neuromorphic systems with neural networks, realized at the software level. Special attention will be dedicated to the difficulties related to the hardware realization of “classic” artificial neuron networks (perceptrons) and drawbacks of their realization at the software level. We will also consider power efficiency of all realized systems. Next, we will discuss several artificial neuron networks realized on memristive devices (perceptrons, spiking neuron networks and reservoir computing) analyzing their advantages and drawbacks. As the successive step, several memristor-based circuits, mimicking learning of living beings will be considered. In particular, a circuit, mimicking classic conditioning (Pavlov’s dog learning), realized using STDP (Spike Timing Dependent Plasticity) algorithm, will be discussed. As the development of this approach, we will discuss ways, towards possible realization of neuro-prostheses, in particular, for the spinal cord damaged patients. Finally, we will consider the possibility of coupling live neurons from cortex by organic memristive devices, playing roles of natural synapses.
10:30
–
11:00
Memristive Devices In Neuromorphic Systems
Victor Erokhin
Director of science @ Institute of Materials for Electronics and Magnetism, Italian National Council of Researches (IMEM-CNR)
12:30
–
13:00
AI and Neurobiology of Deception, Falsehood, and Misleading: How AI Helps Us Understand Ourselves. Insights from Recent Research
Evgenia Alshanskaia
Researcher @ Institute of Cognitive Neurosciences, Personalized Research and Mental Health Group
Neuromorphic computing in spiking neuron-astrocyte networks
Neuromorphic computing, an interdisciplinary field driven by brain-inspired “spiking” computational frameworks, holds the promise of achieving artificial intelligence while significantly lowering the energy consumption of computing platforms. This field, which initially focused on replicating biological neural processes using silicon circuits, has expanded to encompass hardware implementations of algorithms that utilize spike-based encoding and event-driven representations. The human brain has the remarkable ability to store and process phenomenal amounts of information while consuming extremely low energy (around 100 watts, with signal rhythm frequencies not exceeding 200 Hz). Traditionally, neurons were considered the primary signaling cells in the brain. Understanding the mechanisms of information processing in the brain hinges on neuronal interactions. Neurons communicate through electrical impulses (spikes), transmitted via unidirectional connections (synapses). This asynchronous mode of data exchange enables energy-efficient computations, whose mechanisms remain not fully understood. Recently, astrocytes (one of the most abundant types of glial cells) have been found to also be signaling cells in the brain, capable of generating chemical activity impulses in response to neuronal activity. Advancements in experimental methods have revealed that astrocytes influence signal transmission in the neural network and directly participate in information processing, as well as in the development of neurodegenerative diseases and brain aging. Research into intercellular signaling processes in the brain, the development of adequate biologically-plausible mathematical models of these processes, and their hardware implementation represent a significant leap in scientific investigation, aimed at creating conceptually novel neuromorphic devices that mimic the principles of information processing in the brain. Neuromorphic artificial intelligence systems, which are network models comprised of biologically-plausible elements, spiking neurons, and astrocytes, capable of learning according to rules approximating those of natural brain learning, form the algorithmic basis for the functioning of memristive neuroelectronics.
14:00
–
14:30
Neuromorphic computing in spiking neuron-astrocyte networks
Susanna Gordleeva
Professor @ Lobachevsky State University, IT-campus Neymark
Panel: Human Brain vs AI Brain: Can We Design a Truly Human-like AI?
Panel
As artificial intelligence continues to evolve, the question of whether we can create AI systems that replicate human reasoning, comprehension, and understanding has become increasingly important. In this session, we bring together leading experts from the fields of neuroscience, computer science, and entrepreneurship to explore the current state and future potential of AI. We’ll dive into how insights from the human brain, cutting-edge AI models, and practical applications for enterprises are shaping the quest for human-like AI. With major players like OpenAI, Anthropic, and Mistral pushing the boundaries of AI technology, we will discuss whether these advancements can overcome the limitations of current models and unlock the next generation of intelligent systems.
16:00
–
16:45
Panel: Human Brain vs AI Brain: Can We Design a Truly Human-like AI?
Moderator: Tarry Singh, Panelists: Srikanth Ramaswamy & Drazen Orescanin & Jordi Vallverdu
CEO @ Deepkapha & Director @ Neural Circuits Laboratory, Newcastle University & Founder @ Solvership & Researcher @ ICREA
Neuro-Diverse AI: Enhancing Creative Problem Solving through Unconventional Cognitive Strategies and Morphological Designs.
This talk explores the fascinating intersection of artificial intelligence, non-neurotypical reasoning, and morphological computing. While much of AI development has been inspired by the neural systems of the human brain, generative AI offers a unique opportunity to explore non-neurotypical approaches that could revolutionize how we design and utilize intelligent systems. We will examine how generative AI, inspired by diverse cognitive processes, can create more adaptable and robust models. By integrating insights from non-neurotypical reasoning, we can develop AI systems that not only think differently but also bring fresh perspectives and solutions to complex problems. Moreover, we will discuss the burgeoning field of morphological computing, where AI systems are embodied in bioinspired or freely designed physical forms. This approach leverages the principles of natural systems and evolutionary biology, enabling AI to interact with the world in more organic and intuitive ways.
16:45
–
17:15
Neuro-Diverse AI: Enhancing Creative Problem Solving through Unconventional Cognitive Strategies and Morphological Designs.
Jordi Vallverdu
Researcher @ ICREA

Nikola Tesla C
Security-focused AI,
AI & Fraud Prevention
People-Centric Security: From Awareness to Culture
Security-focused AI
Panel with the solution provider and the client
Intermediate
People-Centric Security: From Awareness to Culture is about fixing what’s broken in security training. Traditional programs are dull, forgettable, and don’t change behaviour. This talk covers practical ways to make security second nature for everyone: quick, engaging training, real feedback, and simple tools that actually change the future behaviour. Powering smarter, personalised training and proactive protection.
10:00
–
10:30
People-Centric Security: From Awareness to Culture
Uros Arsenijevic
Product Manager @Seif.aii
Keynote: The High Stakes Security of AI
Keynote
As AI becomes increasingly integrated into our lives, the security of these systems has profound implications for society. AI has proven to be an incredible tool with tremendous benefits, but the risks of malicious actors exploiting vulnerability could be catastrophic. Lets talk about the far-reaching consequences of AI security and the ethical imperative to develop AI systems which are secure by design.
10:30
–
11:15
Keynote: The High Stakes Security of AI
Brian Wagner
CTO @ Revenir
Keynote: AI As The New Hacker Frontier
Keynote
Transformational talk
Beginner
This keynote gives insight into the intersection of AI and cyber security from a hacker’s perspective and what we need to prepare for when it comes to emerging threats. While the business world debates the size of their AI investments and the extent of the latest AI trends, cyber attackers are already harnessing AI for malicious purposes to increase the efficiency and cost-effectiveness of their attacks.
The keynote further addresses the future AI cyber security threat picture and what adaptations will be required for organisations to better prepare for a future that includes AI as part of a comprehensive cyber security strategy.
The keynote further addresses the future AI cyber security threat picture and what adaptations will be required for organisations to better prepare for a future that includes AI as part of a comprehensive cyber security strategy.
11:45
–
12:30
Keynote: AI As The New Hacker Frontier
Tom Van de Wiele
Hacker | Speaker @ Hacker Minded
Keynote: The Future of Cybersecurity in an AI-driven Era
Keynote
Join Francesco for an engaging session on the interconnected worlds of artificial intelligence and cybersecurity. Francesco will illustrate how AI is revolutionizing technology with its immense capabilities while also uncovering the practical cybersecurity risks and challenges it brings, including how threat actors are leveraging it for malicious purposes.
Finally, gain insights into the dual-edge nature of AI—its power both as an ally for defenders and a weapon for attackers—and learn how to effectively leverage it in countering cyber threats. Stay ahead in the AI-driven era by understanding the critical balance between innovation and security, all through engaging stories and practical examples.
Finally, gain insights into the dual-edge nature of AI—its power both as an ally for defenders and a weapon for attackers—and learn how to effectively leverage it in countering cyber threats. Stay ahead in the AI-driven era by understanding the critical balance between innovation and security, all through engaging stories and practical examples.
12:30
–
13:15
Keynote: The Future of Cybersecurity in an AI-driven Era
Francesco Vigo
Director of Engineering in the Office of the Chief Product Officer @ Palo Alto Networks
14:15
–
14:45
How to Leverage AI Securely
Rafah Knight
CEO & Founder @ SecureAI
The Risk of AI adoption and business implication of not using AI
Security-focused AI
– Data problem — who accesses our data, what about shadow AI etc.
– Model testing — AI red teaming and testing is the biggest problem to solve and here SplxAI comes into the game
– Compliance and Security — Why now is always determined by audits and compliance needed to be done and here SplxAI automates those topics
– Model testing — AI red teaming and testing is the biggest problem to solve and here SplxAI comes into the game
– Compliance and Security — Why now is always determined by audits and compliance needed to be done and here SplxAI automates those topics
14:45
–
15:15
The Risk of AI adoption and business implication of not using AI
Kristian Kamber
CEO & Co-Founder @ SplxAI
15:15
–
16:00
Panel: AI in Action: Real-World Solutions for Cybersecurity and Fraud Prevention
Marko Elazar & Branislav Veselinovic & Petar Bajovic & Zarko Kecic
CEO @ Nova defense & Corporate security senior specialist @ UCB Serbia & IT Security Engineer @ Greentube GmbH & CTO @ RNIDS
Vision-Language Models for Fraud Detection
AI & Fraud Prevention
As fraud attacks on ID documents become increasingly diverse, managing separate detection systems for each type can be inefficient and costly. This presentation explores the potential of vision-language models (VLMs) to consolidate multiple fraud detection tasks into a single, powerful system. We’ll outline the key challenges we face and how VLMs offer a scalable, unified solution.
16:30
–
17:00
Vision-Language Models for Fraud Detection
Renat Khizbullin
Senior Machine Learning Engineer @ incode
Guardian – fraud detection using graph analysis
AI & Fraud Prevention
Use case
Intermediate
Guardian uses graph databases, advanced data analytics and ML to detect fraud and suspicious behaviour. The solution leverages a network model and combines data science and visual analytics methods to support users in various domains to more efficiently detect, investigate and prevent fraud. Data is presented in a graph/social network enabling bettwr inspection of the subject relationships and further feature engineering for the underlying ML model. The tool also enables visual inspection and exploration of the network in order for the users to better analyse and inspect suspicious cases.
17:00
–
17:30
Guardian – fraud detection using graph analysis
Tatjana Pejcinovic & Petra Sinka
Data Lead @ True North & Data Scientist @ True North
* this program is not final and is subject to change, full schedule will be available soon