When:
Machine Learning
Week Warsaw
Transform into a data-driven business with machine learning and AI
Experience Machine Learning Week Warsaw
The conference focuses on the practical application of predictive & prescriptive analytics, machine & deep learning, data & text mining, Generative AI & Agentic AI and related topics in business, industry, government, and other verticals, covering analytical and technical as well as economical and organisational questions on strategic and operational levels.
We provide a regional platform for the European data science community to share their success stories and insights with their industry peers. Our attendees are experienced data scientists, analytics managers and AI visionaries from manufacturing, logistics, marketing, e-commerce, finance, insurance, healthcare and many more sectors.
Key focus areas
- Accelerate your transformation into a data-driven business with machine learning and AI.
- Share success stories, in-depth knowledge and advanced insights with industry peers.
- Combine Generative AI and Predictive AI for faster results.
Join us to delve into the practical applications of:
- Moving Fast from Predictive to Generative to Agentic AI
- EverythingOps = DataOps + MLOps + DevOps + StratOps
- From AI Tool Experimentation to AI Business Transformation
- Beyond LLM: Multimodality, (Probabilistic) Reasoning, SML & more
- FAIR Data & Responsible AI: more than GDPR, EU Data Act, EU AI Act
- Practical Applications in Business and Industry
Join the platform for the data science community to share success stories and insights with industry peers from manufacturing, logistics, marketing, e-commerce, financial and many more sectors. Don’t miss this perfect opportunity for in-depth knowledge-sharing, interactive expert discussions, keynotes, case studies, workshops and intensive industry networking.
Deep dive into Generative AI, Predictive AI, and how they can work together to turbocharge your output. Join fellow data scientists, analytics managers and AI visionaries and leave with techniques for bigger wins, broader capabilities and MLOPs success.
Agenda
| 2. October 2025 | ||
| 8:00 – 9:00 | Registration and Breakfast | |
| 9:00 – 9:05 | Opening Moderator: Martin Szugat |
|
| 9:05 – 9:45 | Opening Keynote Rethinking AI: Are We Climbing the Right Hill? OpenAI took an amazing idea – Transformers – and sold it like a dream: ask AI. Now every company has AI-something to sell, and every country is gearing up for the AI race, turning international research into a form of treason. But what if this madness could be our best chance to put aside the over-excitement about connectionism – without throwing away its amazing results – and think about the next class of problems worth solving? It’s time to rethink AI deeply – while we have the money – and start looking for damn good answers, not just good enough. Speaker: Allessandro Confetti |
|
| 9:45 – 10:15 | Sponsored Session | |
| 10:15 – 10:45 | Coffee Break | |
| Track 1 – Case Studies Moderator: Ewa Nowakowska |
Track 2 – Deep Dives Moderator: Martin Szugat |
|
| 10:45 – 11:15 | Accelerating Marketing Delivery: Axel Springer’s Journey to a Unified Next-Gen Automation Platform Nils, Senior Data Engineer, and Justin, Senior Data Scientist, guide you through this Case Study featuring Axel Springer, known for media ventures like the Polish Onet. They illustrate how initially isolated predictive use cases evolved into a next-gen marketing automation platform that accelerates campaign delivery and drives subscriber growth. Justin explains the analytics behind the first ML use cases for better customer targeting, while Nils details critical data integrations that transformed these insights into a unified Customer Data Chip and seamless automation architecture. This session demonstrates the power of data products and serves as a blueprint for marketing enablement. Speakers: Nils Paetsch & Justin Neumann |
Human-in-the-loop: Practical Lessons for Building Comprehensive AI Systems Machine learning models should not live in a vacuum. When a predictive model is exposed to the user to enrich a scientific or industrial workflow, incorporating human feedback is essential for the model to improve over time, overcome distribution shifts, and learn novel phenomena. We address the engineering challenges of building interactive active learning systems on a practical example of large-scale video analysis. Speaker: Miha Garafolj |
| 11:15 – 11:45 | From Eye-Catching to Effective: Personalization That Actually Works in the Mol MOVE Loyalty Program at MOL Group What happens when a data team faces a choice between cutting-edge and commercially effective? Working within Mol MOVE, a fuel retailer loyalty program covering 10 CEE countries, we tested both and share our key learnings from those two projects in this session: 1) a hyperpersonalization project, where we conducted microsegmentation, used GenAI to create data-driven personas, and use them to write personalized message to these microsegments 2) a discount optimization project, where we used a traditional AI approach to determine what would be the optimal discount level for every customer in that specific segment. Speaker: Peter Szabolcs |
|
| 11:45 – 11:50 | Short Break | |
| 11:50 – 12:10 | From Hypotheses to Conversions: Journey to First ML Models at Snaptrip Explore Snaptrip’s path to their first ML Models. How we combined techniques like siamese networks, NLP and ensemble of multiple models in order to achieve conversion improvement on their website. We will describe how we tackled data drift in resource limited environment. You will walk away with description and real tips of practical ways to manage many experiments, and achieving balance between simple and complex approaches. Speaker: Marcin Szymaniuk |
Hijacking Attacks against DP Federated Split Learning Split learning and differential privacy are technologies with growing potential to help with privacy-compliant advanced analytics on distributed datasets. This session will show applying a recent feature space hijacking attack (FSHA) to the learning process of a split neural network enhanced with differential privacy (DP), using a client-side off-the-shelf DP optimizer. It will show the results of the attack and discuss its implications. Speaker: Grzegorz Gawron |
| 12:10 – 12:30 | Causal AI at Scale: Personalization of Push Notifications for Millions of PAYBACK Users In high-scale mobile engagement, sending too many push notifications can alienate users, while too few can limit ROI. We built a causal AI system leveraging GCP to predict the optimal notification frequency for each user based on their behavior and characteristics. By estimating heterogeneous treatment effects through causal forests and validating through A/B testing, we developed a personalized frequency model that maximizes user interest and minimizes fatigue. This talk shares our journey from business constraint to scalable AI-driven solution that now serves millions of users daily Speaker: Dr.Alexander Khachikyan |
|
| 12:30 – 14:00 | Lunch Break | |
| 14:00 – 14:45 | Keynote | |
| 14:45 – 15:00 | Sponsored Session | |
| 15:00 – 15:30 | Coffee Break | |
| 15:30 – 16:30 | Table Discussions You are surrounded by fellow data scientists and there is no better place to discuss and share our common problems – off the record, and in a small group of max. 12 people. These are your people – they understand your situation. Often rated the best part of the event, sharing your problems with like-minded professionals is your path to answers, a little empathy, and a stronger professional network. We will publish the topics and moderators of the roundtable discussions closer to the date. |
Table Discussions You are surrounded by fellow data scientists and there is no better place to discuss and share our common problems – off the record, and in a small group of max. 12 people. These are your people – they understand your situation. Often rated the best part of the event, sharing your problems with like-minded professionals is your path to answers, a little empathy, and a stronger professional network. We will publish the topics and moderators of the roundtable discussions closer to the date. |
| 16:30 – 16:35 | Short Break | |
| 16:35 – 16:50 | From Reactive to Autonomous: Building Agentic AI at Feedly Most AI products today are sophisticated tools; powerful, but passive. At Feedly, we’re transforming threat intelligence from a reactive discipline to a proactive partnership with AI. This talk traces our journey from RSS feeds to agentic AI workflows, sharing real lessons from building and deploying AI features used by thousands of security teams. I’ll demonstrate how we evolved from keyword matching to semantic AI Feeds, then added generative AI capabilities that analyze, summarize, and create intelligence reports. More importantly, I’ll share our vision for the next frontier: agentic AI that acts as a trusted security colleague; continuously hunting threats, understanding organizational context, and taking autonomous actions. Speaker: Farah Ayadi |
Advanced Prompt Engineering and Beyond: Ingredients Needed for Working LLM Based Products We start with examples of advanced prompt engineering. Then we will show limitations of various approaches and iterate through retrieval techniques and verification. We’ll dive into Hyde and Colbertv2 for ReRank and better retrieval, and Chain of Thought for reasoning. Learn how LLMs can act as judges, balancing fairness and precision. We will also explore when it makes sense to use PEFT. We’ll tackle challenges like sneaky user behavior and testing non-deterministic outputs. Speaker: Marcin Szymaniuk |
| 16:50 – 17:05 | TBA | |
| 17:05 – 17:10 | Short Break | |
| 17:10 – 17:25 | How Not to Get Burned: Security Lessons from Real-World LLM Deployments Many LLM projects fail to deliver secure, reliable results due to overlooked risks like data leaks, prompt injection, and poor governance. This session unpacks common security pitfalls through 10 real-world cases from finance to healthcare—then offers hands-on strategies for safer LLM design, deployment, and oversight. Learn how to align innovation with safety, compliance, and trust at every stage of the AI lifecycle. Speaker: Dawid Pacholczyk |
TBA |
| 17:25 – 17:45 | TBA | |
| 17:45 – 17:55 | Wrap Up | |
| 17:55 | Reception in the Exhibit Hall | |
Why attend?
Machine Learning Week Warsaw is where real AI/ML gets done. Here’s what you’ll gain:
- Global format, local stage
For the first time in Poland, experience the renowned Machine Learning Week format – developed by Rising Media and trusted by the European AI community. - 100% practical content, 0% fluff
Every session is a case study. Learn directly from real AI/ML implementations – what worked, what didn’t, and what’s next. - Network with real practitioners
Meet data scientists, ML engineers, and AI leaders who speak your language – no hype, just real talk and shared experience. - Topics that matter right now
From LLMs and Generative AI to RAG, ML Ops, EverythingOps, and the EU AI Act – get insights that are timely, relevant, and actionable. - One powerful day, maximum value
No week-long commitment needed. Just one day packed with cutting-edge content, fresh ideas, and high-value connections.
Machine Learning
Week Warsaw
Keynote Speaker
Alessandro Confetti
European Data & AI lead for Health Care,
Life Science, and Tech
ThoughtWorks
Alessandro started writing software when he was fourteen, and he has never stopped since then. After studying philosophy, logic, and working with healthcare data for his entire career, he joined ThoughtWorks in 2017 helping large organizations make sense of their legacy software and data. Since 2019, he has been an adjunct professor at Politecnico of Milan, and from 2022 also a Design School advisory board member. Currently, he’s ThoughtWorks European Data & AI lead for Health Care, Life Science, and Tech.
Program Chair
Martin Szugat
Founder & Managing Director
Datentreiber
Founder and managing director of Datentreiber, a data strategy consultancy supporting companies to digitally transform to a data-driven business. He invented the Data Strategy Design method and provides a free Data Strategy Designkit. Prior to Datentreiber, Martin Szugat was managing director of SnipClip, an agency for social media marketing & analytics. With a degree in bioinformatics, he has conducted research in machine learning and worked as a freelance software developer, consultant & author. Since 2014, he is responsible for the Machine Learning Week (formerly known as Predictive Analytics World & Deep Learning World) conference in Europe as program director.
Speakers
Miha Garafolj
Senior ML Engineer
Merantix Momentum
Miha’s academic background is in Financial mathematics, in which he obtained a MSc from the University of Ljubljana. Already during studies he got pulled into applied ML, first using ML and other quantitative algorithms to help solve financial portfolio optimization problems, and later transitioning into computer vision. For more than two years, together with his colleagues in Merantix Momentum, he specializes in data-centric AI and designing human in the loop systems for their clients.
Alexander Khachikyan
Principal AI & Data Scientist
PAYBACK
Alexander leads the development and scaling of AI-driven products generating multimillion-euro impact across PAYBACK’s digital ecosystem. He holds a Ph.D. from the Max Planck Institute and teaches Machine Learning and Deep Learning at the Munich University of Applied Sciences.
Justin Neumann
Senior Data Scientist
Axel Springer National Media & Tech
Justin Neumann works as a Senior Data Scientist at Axel Springer National Media & Tech, the digital powerhouse of Europe’s leading publisher. He has architected and developed machine-learning solutions for customer targeting and marketing automation, accelerating marketing delivery and driving subscriber growth. Channeling his passion for technology, he continually refines data products that enable marketing teams to succeed. He holds a Master of Science in Predictive Analytics from Northwestern University.
Ewa Nowakowska
Partner and Leader of the AI Data Science Team
EY Poland
Ewa leads the AI Data Science Team at EY Poland and is a Partner at the firm. She holds a PhD in Computer Science from the Polish Academy of Sciences and degrees in Mathematics and Psychology from the University of Warsaw. With over 15 years of experience, she has worked on AI and machine learning projects in Europe and the US, supporting both global initiatives and client solutions. At EY, she focuses on building AI systems tailored to business needs and fostering collaboration across teams. She is also active in AI education and mentoring, particularly supporting young talent and women in tech.
Nils Paetsch
Senior Data Engineer
Axel Springer National Media & Tech
Nils Paetsch, Senior Data Engineer, and Justin Neumann, Senior Data Scientist, work at Axel Springer National Media & Tech, the digital powerhouse of Europe’s largest publisher. Together, they drive subscriber growth and empower stakeholders and marketing teams to expand media businesses both in Europe and internationally.
Szabolcs Péter
Lead Data Scientist
Digital Factory (Member of MOL Group)
Meet Peter, a seasoned data scientist with a knack for unraveling insights. Beyond his technical prowess, he’s known for his passion for teaching, having enlightened hundreds on the intricacies of data science and data visualization. Currently, he serves as a Lead Data Scientist and trailblazer in AI solutions at Digital Factory, where he continues to push the boundaries of innovation. Join him at the conference for a blend of expertise, humor, and a shared love for both data and nature.
Arkadiusz Slowik
Data Scientist
PAYBACK
Mathematics graduate from the University of Warsaw with 2 years of end-to-end ML experience at PAYBACK. Multi-time hackathon winner and former enthusiast. Speaker and roundtable moderator at tech conferences, passionate about sharing practical insights and fostering discussion in the data science and machine learning community
Machine Learning
Week Warsaw
Venue
The Tides,
Wioślarska 8, Warsaw
Why sponsor?
Machine Learning Week Warsaw offers unmatched visibility to decision-makers driving AI in the CEE region:
- Direct access to AI/ML decision-makers from Poland and the CEE region
Connect directly with the people who matter – those leading AI and data initiatives across top enterprises in Central and Eastern Europe. - Real presence in the spotlight
Get more than just a logo. Lead a breakout session, present a real-world case study, or run a hands-on workshop – and showcase your expertise in action. - Warsaw – the rising AI hub of Central Europe
Position your brand in the heart of one of Europe’s fastest-growing tech regions. Warsaw is a natural launchpad for business expansion across CEE. - International standards, local impact
Co-organized with Rising Media (creators of ML Week Europe), the event delivers the same expert-led program and content quality trusted across Berlin, Munich, and London. - Your target audience: companies actively investing in AI
Meet senior practitioners actively building and scaling real-world AI – ready to engage, evaluate, and partner.
Get in front of the people shaping AI in Central Europe. Connect with AI leaders in Poland and beyond.
Sponsor Opportunities
Hosts
Agenda
| 2. October 2025 | ||
| 8:00 – 9:00 | Registration and Breakfast | |
| 9:00 – 9:05 | Opening Moderator Martin Szugat |
|
| 9:05 – 9:45 | Opening Keynote Rethinking AI: Are We Climbing the Right Hill? OpenAI took an amazing idea – Transformers – and sold it like a dream: ask AI. Now every company has AI-something to sell, and every country is gearing up for the AI race, turning international research into a form of treason. But what if this madness could be our best chance to put aside the over-excitement about connectionism – without throwing away its amazing results – and think about the next class of problems worth solving? It’s time to rethink AI deeply – while we have the money – and start looking for damn good answers, not just good enough. Speaker: Allessandro Confetti |
|
| 9:45 – 10:15 | Sponsored Session | |
| 10:15 – 10:45 | Coffee Break | |
| Track 1 – Case Studies Moderator: Ewa Nowakowska |
Track 2 – Deep Dives Moderator: Martin Szugat |
|
| 10:45 – 11:15 | Accelerating Marketing Delivery: Axel Springer’s Journey to a Unified Next-Gen Automation Platform Nils, Senior Data Engineer, and Justin, Senior Data Scientist, guide you through this Case Study featuring Axel Springer, known for media ventures like the Polish Onet. They illustrate how initially isolated predictive use cases evolved into a next-gen marketing automation platform that accelerates campaign delivery and drives subscriber growth. Justin explains the analytics behind the first ML use cases for better customer targeting, while Nils details critical data integrations that transformed these insights into a unified Customer Data Chip and seamless automation architecture. This session demonstrates the power of data products and serves as a blueprint for marketing enablement. Speakers: Nils Paetsch & Justin Neumann |
Human-in-the-loop: Practical Lessons for Building Comprehensive AI Systems Machine learning models should not live in a vacuum. When a predictive model is exposed to the user to enrich a scientific or industrial workflow, incorporating human feedback is essential for the model to improve over time, overcome distribution shifts, and learn novel phenomena. We address the engineering challenges of building interactive active learning systems on a practical example of large-scale video analysis. Speaker: Miha Garafolj |
| 11:15 – 11:45 | From Eye-Catching to Effective: Personalization That Actually Works in the Mol MOVE Loyalty Program at MOL Group What happens when a data team faces a choice between cutting-edge and commercially effective? Working within Mol MOVE, a fuel retailer loyalty program covering 10 CEE countries, we tested both and share our key learnings from those two projects in this session: 1) a hyperpersonalization project, where we conducted microsegmentation, used GenAI to create data-driven personas, and use them to write personalized message to these microsegments 2) a discount optimization project, where we used a traditional AI approach to determine what would be the optimal discount level for every customer in that specific segment. Speaker: Peter Szabolcs |
|
| 11:45 – 11:50 | Short Break | |
| 11:50 – 12:10 | From Hypotheses to Conversions: Journey to First ML Models at Snaptrip Explore Snaptrip’s path to their first ML Models. How we combined techniques like siamese networks, NLP and ensemble of multiple models in order to achieve conversion improvement on their website. We will describe how we tackled data drift in resource limited environment. You will walk away with description and real tips of practical ways to manage many experiments, and achieving balance between simple and complex approaches. Speaker: Marcin Szymaniuk |
Hijacking Attacks against DP Federated Split Learning Split learning and differential privacy are technologies with growing potential to help with privacy-compliant advanced analytics on distributed datasets. This session will show applying a recent feature space hijacking attack (FSHA) to the learning process of a split neural network enhanced with differential privacy (DP), using a client-side off-the-shelf DP optimizer. It will show the results of the attack and discuss its implications. Speaker: Grzegorz Gawron |
| 12:10 – 12:30 | Causal AI at Scale: Personalization of Push Notifications for Millions of PAYBACK Users In high-scale mobile engagement, sending too many push notifications can alienate users, while too few can limit ROI. We built a causal AI system leveraging GCP to predict the optimal notification frequency for each user based on their behavior and characteristics. By estimating heterogeneous treatment effects through causal forests and validating through A/B testing, we developed a personalized frequency model that maximizes user interest and minimizes fatigue. This talk shares our journey from business constraint to scalable AI-driven solution that now serves millions of users daily Speaker: Dr.Alexander Khachikyan |
|
| 12:30 – 14:00 | Lunch Break | |
| 14:00 – 14:45 | Keynote | |
| 14:45 – 15:00 | Sponsored Session | |
| 15:00 – 15:30 | Coffee Break | |
| 15:30 – 16:30 | Table Discussions You are surrounded by fellow data scientists and there is no better place to discuss and share our common problems – off the record, and in a small group of max. 12 people. These are your people – they understand your situation. Often rated the best part of the event, sharing your problems with like-minded professionals is your path to answers, a little empathy, and a stronger professional network. We will publish the topics and moderators of the roundtable discussions closer to the date. |
Table Discussions You are surrounded by fellow data scientists and there is no better place to discuss and share our common problems – off the record, and in a small group of max. 12 people. These are your people – they understand your situation. Often rated the best part of the event, sharing your problems with like-minded professionals is your path to answers, a little empathy, and a stronger professional network. We will publish the topics and moderators of the roundtable discussions closer to the date. |
| 16:30 – 16:35 | Short Break | |
| 16:35 – 16:50 | From Reactive to Autonomous: Building Agentic AI at Feedly Most AI products today are sophisticated tools; powerful, but passive. At Feedly, we’re transforming threat intelligence from a reactive discipline to a proactive partnership with AI. This talk traces our journey from RSS feeds to agentic AI workflows, sharing real lessons from building and deploying AI features used by thousands of security teams. I’ll demonstrate how we evolved from keyword matching to semantic AI Feeds, then added generative AI capabilities that analyze, summarize, and create intelligence reports. More importantly, I’ll share our vision for the next frontier: agentic AI that acts as a trusted security colleague; continuously hunting threats, understanding organizational context, and taking autonomous actions. Speaker: Farah Ayadi |
Advanced Prompt Engineering and Beyond: Ingredients Needed for Working LLM Based Products We start with examples of advanced prompt engineering. Then we will show limitations of various approaches and iterate through retrieval techniques and verification. We’ll dive into Hyde and Colbertv2 for ReRank and better retrieval, and Chain of Thought for reasoning. Learn how LLMs can act as judges, balancing fairness and precision. We will also explore when it makes sense to use PEFT. We’ll tackle challenges like sneaky user behavior and testing non-deterministic outputs. Speaker: Marcin Szymaniuk |
| 16:50 – 17:05 | TBA | |
| 17:05 – 17:10 | Short Break | |
| 17:10 – 17:25 | How Not to Get Burned: Security Lessons from Real-World LLM Deployments Many LLM projects fail to deliver secure, reliable results due to overlooked risks like data leaks, prompt injection, and poor governance. This session unpacks common security pitfalls through 10 real-world cases from finance to healthcare—then offers hands-on strategies for safer LLM design, deployment, and oversight. Learn how to align innovation with safety, compliance, and trust at every stage of the AI lifecycle. Speaker: Dawid Pacholczyk |
TBA |
| 17:25 – 17:45 | TBA | |
| 17:45 – 17:55 | Wrap Up | |
| 17:55 | Reception in the Exhibit Hall | |