






The origins of the Data Science Summit date back to 2014, when the first Data Science Warsaw community meeting was held. It was the popularity of these periodic meetups that contributed to the decision to organize the first edition of the conference, and its success to subsequent editions. Currently, DSS is the largest Polish community focused around *data* field.
The DSS is the CEE region's largest *data* conference, bringing together more than 2,500 people from different backgrounds and areas each year. The conference is accompanied by the industry's Data Science Expo.
A special edition of the conference fully dedicated to the field of machine learning and artificial intelligence. It attracts 1,500+ attendees every year. The conference is accompanied by the industry's Data Science Expo.
This is a unique conference in Poland dedicated to developers and other software professionals who want to stay on top of trends and implement modern AI-based solutions.
DSW is the largest data science community in Poland, based in Warsaw, Poland. It is an informal non-profit organization working to exchange ideas and knowledge about data science, data engineering and so-called artificial intelligence. The purpose of its existence is to discuss tools, technology and business opportunities related to collecting, processing and visualization of data, as well as machine learning and deep learning.
Data Scientists
AI / ML Researchers
AI / ML Engineers
MLOps / DataOps
Cloud Engineers
Data Analysts
Big Data / Data Platforms Engineers
Business Intelligence Specialists
Managers in IT / Data / Compliance areas
CDO, CTO, CIO
...and anyone else working in the areas of data analysis, processing and visualization
Data warehouses, database systems integration, stream processing, and DataOps.
Model scalability in cloud architectures (AWS, Azure, GCP) and inference on target devices (On-device AI / SLMs).
Production architectures for AI systems, continuous concept drift monitoring, and model lifecycle management.
Advanced development tools (Python, R, Scala) and computational performance optimization.
LLM architectures, Retrieval-Augmented Generation (RAG) systems, and hallucination mitigation techniques.
Image analysis, convolutional neural networks, and automated pattern detection.
Time series, survival analysis, and the transition from purely correlational modeling toward causal inference.
Reinforcement Learning, autonomous systems, and Agentic AI.
Scaling AI projects from Proof of Concept to production deployment. Business model redesign, operational process optimization, transaction cost analysis of AI adoption, and evaluation metrics (ROI/TCO) for machine learning investments.
Structuring analytical competencies within organizations. Managing multidisciplinary Data Science teams, project lifecycle management (agile in analytics), employee reskilling/upskilling strategies, the role of AI Champions, and change management.
Treating data as a business asset (Data as a Product). Building organizational data maturity, democratizing data access (Data Mesh/Data Fabric from a management perspective), Master Data Management (MDM), and maintaining data quality for ML models.
Operationalizing regulatory requirements in commercial environments. Implementing model auditability standards and the AI Act from a business continuity and risk management perspective.
Rigorous applications of analytics in medicine.
Modeling spatial and sensor-based (IoT) data.
Advanced executive reporting, model interpretability (XAI), and decision-support visualization.
Security of analytical ecosystems, anomaly detection, and fraud prevention in the financial sector. The role of AI in AML, fraud detection, and cybersecurity for business processes.






* for selected packages
At the event you will meet and listen to the best specialists from Poland and abroad.












Check Agenda 2026 - Selected by practitioners, the most interesting speeches on trends, latest solutions and production experiences.
19:00 - 00:00
Kandela Bar (Elektrownia Powiśle)
08:00 - 08:30
PGE Narodowy
08:30 - 09:00
La Cantina (P1)
Oficjalne Otwarcie | Official Opening
09:00 - 09:35
La Cantina (P1)
[Special Guest] Fireside Chat
09:45 - 10:20
La Cantina (P1)
Data science methods that drive localization
10:35 - 11:10
La Cantina (P1)
Silicon Lifeforms. Humans in the loop but who is in control?
11:20 - 11:55
La Cantina (P1)
Make The Difference: AI ROI
12:10 - 12:45
La Cantina (P1)
From AI North Star Strategy to Deployed AI Agents in Comarch Products
12:55 - 13:30
La Cantina (P1)
[Track Keynote] Bielik v3.1. Prezentacja i omówienie powstawania modelu językowego
13:35 - 14:20
La Cantina (P1)
[MAIN STAGE SPECIAL] Modele system-one - od generowania do decydowania. Nowa era AI w biznesie
14:20 - 14:55
La Cantina (P1)
Using Local Language Models to Assess the Quality of Narrative Data in Operational Risk
15:05 - 15:40
La Cantina (P1)
Beyond the model: Designing AI systems that solve real problems. Three practical lessons and what they mean for the evolving role of the data scientist.
15:55 - 16:30
La Cantina (P1)
Embeddings for Earth: How Foundation Models Are Reshaping GeoAI
16:40 - 17:15
La Cantina (P1)
Automating the process of scientific discovery in practice: a deep dive into Molecule.one and OpenAI near-autonomous discovery in organic chemistry
17:25 - 18:00
La Cantina (P1)
Energy storage – estimating investment profitability
09:00 - 09:35
Paryż A (P2-L1)
927 odpowiedzi, dwie fale, rok różnicy: jak mierzyliśmy adopcję AI w polskim software house i co z tego wynika
09:45 - 10:20
Paryż A (P2-L1)
[Special Guest] Jak zbudowaliśmy zespół data science, który nie obiecuje, tylko dowozi?
10:35 - 11:10
Paryż A (P2-L1)
Same Data & AI Roadmap, Different Stage: How LOT Airlines, Colonnade Insurance and Haddad Brands Modernized Their Data Platforms
11:20 - 11:55
Paryż A (P2-L1)
From Legacy to AI-Ready: Spec-Driven Data Modernization with EPAM
12:10 - 12:45
Paryż A (P2-L1)
What Agentic AI taught us about Implementing Agentic AI Solutions
12:55 - 13:30
Paryż A (P2-L1)
[PANEL DYSKUSYJNY] Skalowanie AI w dużych organizacjach - porównanie perspektyw
13:30 - 14:20
VIP Zone (Loża Książęca)
14:20 - 14:55
Paryż A (P2-L1)
Od obietnic AI do realnego ROI: Jak sprzedać analitykę konwersacyjną dyrektorowi finansowemu
15:05 - 15:40
Paryż A (P2-L1)
Kto ma czerwony przycisk? O kontroli, autonomii AI na produkcji.
15:55 - 16:30
Paryż A (P2-L1)
Od dokumentów do wiedzy. Kulisy budowy bazy wiedzy AI w dużej organizacji
16:40 - 17:15
Paryż A (P2-L1)
The Hidden Failure Point in AI Projects: Externalizing Tacit Knowledge Before Scaling from PoC to Production
17:25 - 18:00
Paryż A (P2-L1)
ISO 9001 without a consultancy: how AI agents built – and now run – a certified quality management system
09:00 - 09:35
Paryż B (P2-L2)
From Policy to Practice: Operationalizing EU AI Act Article 17 Quality Management System in Regulated Industries
09:45 - 10:20
Paryż B (P2-L2)
Monitorowanie postępu technologicznego - jak zaplanować migrację na PQC zgodnie z RODO.
10:35 - 11:10
Paryż B (P2-L2)
AI works exactly as it should. That’s the problem. Why AI adoption exposes flaws that have remained hidden in data for years
11:20 - 11:55
Paryż B (P2-L2)
Why should Data Governance be AI's best friend
12:10 - 12:45
Paryż B (P2-L2)
Everyone Wanted LLMs. Production Wanted FastText: A Pragmatic Approach to Million-Scale Categorization
12:55 - 13:30
Paryż B (P2-L2)
Sequence-Based or Tabular models for Next Best Action Engines? Lessons Learned from the PwC Hackaton
13:30 - 14:20
1 i 2 piętro
14:20 - 14:55
Paryż B (P2-L2)
Everyone wants more AI. But who is going to govern it?
15:05 - 15:40
Paryż B (P2-L2)
When AI Meets Compliance: Building an LLM Agent for Financial KYC Process
15:55 - 16:30
Paryż B (P2-L2)
Six safety gates across 13 stages: the latency and review cost of a governed content generation pipeline
16:40 - 17:15
Paryż B (P2-L2)
Autonomous Enterprise: Reference Architecture for Multi-Agent AI Platforms
17:25 - 18:00
Paryż B (P2-L2)
Where Did This Value Come From? Governing AI-Assisted Data Enrichment in Systems of Record
09:00 - 09:35
Warszawa (P2-L5)
[TRACK KEYNOTE] BI isn't dead. It melted.
09:45 - 10:20
Warszawa (P2-L5)
Koniec dashboardów? Jak w 30 minut zacząć zadawać pytania własnym danym
10:35 - 11:10
Warszawa (P2-L5)
Let’s Get Ready to Rumble: Power BI + Copilot vs. Databricks AI/BI + Genie
11:20 - 11:55
Warszawa (P2-L5)
[SPECIAL GUEST] Power BI + MCP: niech agent zrobi model za Ciebie
12:10 - 12:45
Warszawa (P2-L5)
Czy agenci będą oglądać dashboardy za nas? Dwugłos o roli AI w BI.
12:55 - 13:30
Warszawa (P2-L5)
[PANEL EKSPERCKI] Przyszłość Data & Analytics
13:30 - 14:20
1 i 2 piętro
14:20 - 14:55
Warszawa (P2-L5)
Blank Canvas, Live Data: Painting Dashboards with Snowflake CoCo Agent
15:05 - 15:40
Warszawa (P2-L5)
AI in BI to boost UA
15:55 - 16:30
Warszawa (P2-L5)
Czy PowerPoint już umarł?
16:40 - 17:15
Warszawa (P2-L5)
Prawdziwa podróż klienta wygląda inaczej, niż myślisz. Jak wykorzystujemy process mining do analizy ścieżek klienta w kontaktach z bankiem
17:25 - 18:00
Warszawa (P2-L5)
Kiedy atrybucja kłamie: jak przebudowaliśmy pomiar efektywności marketingu z wykorzystaniem BigQuery i Marketing Mix Modeling
09:00 - 09:35
Barcelona (P2-L4)
Data Platform as a Product, czyli gdzie kończy się pipeline, a zaczyna kontrakt
09:45 - 10:20
Barcelona (P2-L4)
Open format isn't enough: Building the interoperable data stack
10:35 - 11:10
Barcelona (P2-L4)
RAG to za mało: jak mierzyć jakość danych, zanim asystent AI zarekomenduje decyzję biznesową
11:20 - 11:55
Barcelona (P2-L4)
The Rise of the "LLM Data Engineer": Preparing Data for Generative AI at Scale
12:10 - 12:45
Barcelona (P2-L4)
The Messy Middle of Industrial Data: From Production Systems to Business Insight
12:55 - 13:30
Barcelona (P2-L4)
Self-Service at Scale: The Data Platform Powering Netflix's ML Feature Pipelines
13:30 - 14:20
1 i 2 piętro
14:20 - 14:55
Barcelona (P2-L4)
Implementing Data Hubs in Large Organizations: Lessons from the Field
15:05 - 15:40
Barcelona (P2-L4)
One user vector to rule them all - how we put ~200 behavioural features online
15:55 - 16:30
Barcelona (P2-L4)
Snowflake CoCo — jak AI agent zmienia codzienną pracę data engineera
16:40 - 17:15
Barcelona (P2-L4)
Where the RAM Goes: Sizing a Billion-Vector Index from 3.3 TB Down to 65 GB
17:25 - 18:00
Barcelona (P2-L4)
Zero-ETL czy tylko mniej ETL? Mirroring, Copy Job i Pipelines w Microsoft Fabric
14:20 - 14:55
Paryż C (P2-L3)
[Special Guest] Evaluation and red teaming of models and agents
15:05 - 15:40
Paryż C (P2-L3)
Optimizing Ultralytics YOLO26: From 41 FPS up to 922 FPS in 30 minutes
15:55 - 16:30
Paryż C (P2-L3)
Scaling Enterprise AI
16:40 - 17:15
Paryż C (P2-L3)
[Track Keynote] Scaling the Brain: ML-Driven Vertical Autoscaling for GKE Control Plane
17:25 - 18:00
Paryż C (P2-L3)
FinAIOps, using AI to predict, explain and optimise cloud spent
09:00 - 09:35
Paryż C (P2-L3)
Fairness to wybór, nie metryka. Wyzwania debiasingu modeli.
09:45 - 10:20
Paryż C (P2-L3)
Building Nation-Wide Device Identity: The Algorithms Behind Banking Fingerprinting
10:35 - 11:10
Paryż C (P2-L3)
Wielowarstwowy system antyfraudowy
11:20 - 11:55
Paryż C (P2-L3)
Holistyczna analiza klienta pod kątem AML z wykorzystaniem narzędzi AI
12:10 - 12:45
Paryż C (P2-L3)
12:55 - 13:30
Paryż C (P2-L3)
Neural challenger model for investment risk profiling
13:30 - 14:20
1 i 2 piętro
09:00 - 09:35
Londyn B (P2-R2)
Uncovering Clinical AI Deployment Risks with Generative Interventional Attributions
09:45 - 11:10
Londyn B (P2-R2)
[PANEL EKSPERCKI] Jak używać danych medycznych? Rola RCMC w ekosystemie zdrowia
11:20 - 11:55
Londyn B (P2-R2)
MolPILE - large-scale, diverse dataset for molecular representation learning
12:10 - 12:45
Londyn B (P2-R2)
Beyond the Ranking: Investigating the Gene Behind a GWAS Association
12:55 - 13:30
Londyn B (P2-R2)
Modality Fusion in Drug Development
13:30 - 14:20
1 i 2 piętro
14:20 - 14:55
Londyn B (P2-R2)
We shipped a Regulatory Intelligence Deep Agent to 2,500 users. Here's What ML engineering doesn't prepare you for.
15:05 - 15:40
Londyn B (P2-R2)
Hearing diagnosis in 1.5 seconds: wavelet features and classical ML on 5,652 clinical ABR signals
15:55 - 16:30
Londyn B (P2-R2)
22 Hits to the Right Side of the Head: Turning Raw Impact Telemetry into Decisions Combat Athletes Trust; Without Overclaiming
16:40 - 17:15
Londyn B (P2-R2)
FertiPredicta: predykcja skuteczności IVF przed i w kolejnych cyklach leczenia
17:25 - 18:00
Londyn B (P2-R2)
Beyond Model Accuracy: Turning Healthcare Infection Risk Prediction into Real Process Improvement
09:00 - 09:35
Londyn C (P2-R3)
[SPECIAL GUEST] Very difficult is not that difficult anymore: what you need in place to build a custom high-performance solver with AI agents
09:45 - 10:20
Londyn C (P2-R3)
Pół miliarda decyzji cenowych dziennie - produkcyjny ML w optymalizacji aukcji reklamowych
10:35 - 11:10
Londyn C (P2-R3)
Od dokumentu do danych - ekstrakcja informacji finansowych ze sprawozdań z KRS z użyciem AI
11:20 - 11:55
Londyn C (P2-R3)
When Simpler Models Win: Building a Practical Categorisation & Product-Matching Pipeline
12:10 - 12:45
Londyn C (P2-R3)
One Config-Millions of Events. Building a proprietary solution to configure, deploy, and monitor large-scale experiments
12:55 - 13:30
Londyn C (P2-R3)
How will quantum computers change the way we work with AI and Machine Learning?
13:30 - 14:20
1 i 2 piętro
14:20 - 14:55
Londyn C (P2-R3)
[PANEL] AI-assisted data science: implications for team structure, role descriptions and everyday work
15:05 - 15:40
Londyn C (P2-R3)
Embedding physics priors into machine learning: lessons on modeling process dynamics from a steam-saving project in food manufacturing.
15:55 - 16:30
Londyn C (P2-R3)
Odporni na hype: Dlaczego w optymalizacji przestrzeni Empiku postawiliśmy na klasyczną ekonometrię i modele mieszane?
16:40 - 17:15
Londyn C (P2-R3)
What Happens When Time Series Foundation Models Leave the Benchmark and Enter the Real World
17:25 - 18:00
Londyn C (P2-R3)
Representation learning approach to taxonomy analysis
14:20 - 14:55
Londyn A (P2-R1)
Learning to Rank Promotional Artwork with a Vision-Language Model
15:05 - 15:40
Londyn A (P2-R1)
[TOPIC SOON: SOFTSERVE]
15:55 - 16:30
Londyn A (P2-R1)
Don't Make the Network Rediscover Anatomy: Computer Vision for Very Early-Stage Lung Cancer in LDCT
16:40 - 17:15
Londyn A (P2-R1)
[Special Guest] Beyond the Model: Practical Considerations for Deploying AI-Driven Robots in the Real World
17:25 - 18:00
Londyn A (P2-R1)
AI z lotu ptaka
09:00 - 09:35
Londyn A (P2-R1)
Can agentic AI bridge the gap between Earth Observation and emergency response?
09:45 - 10:20
Londyn A (P2-R1)
Dostawa jutro czy za dwa dni? Geolokalizacyjny ML w czasie rzeczywistym przy 100 000 RPS na Allegro
10:35 - 11:10
Londyn A (P2-R1)
Demokratyzacja obserwacji Ziemi: Wolny GIS i otwarte dane satelitarne w społecznej analizie pożarów
11:20 - 11:55
Londyn A (P2-R1)
[Track Keynote] When the 95% Accuracy Is Not Enough: Explaining, Verifying, and Validating AI Models for Space Applications
12:10 - 12:45
Londyn A (P2-R1)
Wszystko, czego baliście się zapytać o OpenStreetMap, a moze się przydać w pracy.
12:55 - 13:30
Londyn A (P2-R1)
[Special Guest] [Panel]
13:30 - 14:20
1 i 2 piętro
09:00 - 09:35
Rzym (P2-R5)
GraphRAG in Practice: Five Design Patterns for Reliable Knowledge Extraction
09:45 - 10:20
Rzym (P2-R5)
Retrofitting Agentic RAG: The Missing NER Layer for Deterministic Recall. Finding and matching products across domains and millions of pages a week
10:35 - 11:10
Rzym (P2-R5)
[Track Keynote] Powering the Map of the Law: Distilling LLMs for Scalable Legal Act Processing
11:20 - 11:55
Rzym (P2-R5)
Prune the experts you don't serve: targeted shrinking of Mixture-of-Experts LMs
12:10 - 12:45
Rzym (P2-R5)
Jak wycisnąć maksimum z dostępnej infrastruktury? Od optymalizacji agenta głosowego do inferencji na GPU
12:55 - 13:30
Rzym (P2-R5)
When Gen AI takes advantage over traditional techniques - and when it does
13:30 - 14:20
1 i 2 piętro
14:20 - 14:55
Rzym (P2-R5)
[PANEL] Czy potrafimy dobrze oceniać AI?
15:05 - 15:40
Rzym (P2-R5)
How to Evaluate AI When There Is No Correct Answer: 3 CVs, 600 Interviews, and One Hidden Prompt
15:55 - 16:30
Rzym (P2-R5)
[Special Guest] Writing in a Genre Is a Specialised Task: LLM Domain and Task Adaptation for Polish Official Writing
16:40 - 17:15
Rzym (P2-R5)
From Assistants to Executable Workflows: A New Paradigm for Industrial AI
17:25 - 18:00
Rzym (P2-R5)
Myśleć, mówiąc: trójstrumieniowa architektura dla niskolatencyjnych agentów głosowych.
09:00 - 09:35
Amsterdam (P2-R4)
Agentic Memory at Scale: a practical guide
09:45 - 10:20
Amsterdam (P2-R4)
When AI Agents Enter the Physical World: Humanoids, Robots and the Rise of Dark Factories
10:35 - 11:10
Amsterdam (P2-R4)
AI on device / AI beyond the cloud
11:20 - 11:55
Amsterdam (P2-R4)
Can Genie Code Productionize an AI Agent? A Controlled Engineering Experiment
12:10 - 12:45
Amsterdam (P2-R4)
Semantic Analytics of AI Interactions
12:55 - 13:30
Amsterdam (P2-R4)
Elevating conversational AI with domain expertise via Chemical Expert Model Context Protocol (MCP) – Merck KGaA case study
13:30 - 14:20
1 i 2 piętro
14:20 - 14:55
Amsterdam (P2-R4)
[Special Guest] How to Break an AI Agent? - Red Teaming Agentic AI Applications.
15:05 - 15:40
Amsterdam (P2-R4)
Ludzie i agenty zmyślają - jak pomóc agentom? Pentesterzy i detektywi AI testują hipotezy.
15:55 - 16:30
Amsterdam (P2-R4)
AI Agents Meet Predictive Modeling: Didn't See That Coming
16:40 - 17:15
Amsterdam (P2-R4)
Try. Fail. Codify. Repeat. — 100 Days of AI as an Engineering Manager
17:25 - 18:00
Amsterdam (P2-R4)
Czy no/low-code automatyzacja jaką znamy, jeszcze ma sens w czasach AI?
10:35 - 11:10
P1-139
AI w organizacjach
11:20 - 11:55
P1-139
AI at scale in healthcare: how to protect models and data against attacks
12:10 - 12:45
P1-139
Lokalne modele - LLMy - wrażenia, takeaways z wystawienia lokalnych modeli
12:55 - 13:30
P1-139
Quo vaditis, automata: Autonomous research and coding agents
14:20 - 14:55
P1-139
Od dojrzałego BI do dojrzałego AI - jak firma z uporządkowaną analityką robi kolejny krok
15:05 - 15:40
P1-139
15:55 - 16:30
P1-139
Gdzie jestem, gdzie idę? Rola analityka BI w organizacji
16:40 - 17:15
P1-139
Agentic AI w sektorze finansowym
09:40 - 09:55
Networking Zone
Ask a Speaker: Data Platform as a Product, czyli gdzie kończy się pipeline, a zaczyna kontrakt
09:40 - 09:55
Networking Zone
Ask a Speaker: [SPECIAL GUEST] Very difficult is not that difficult anymore: what you need in place to build a custom high-performance solver with AI agents
09:40 - 09:55
Networking Zone
Ask a Speaker: Fairness to wybór, nie metryka. Wyzwania debiasingu modeli.
09:40 - 09:55
Networking Zone
Ask a Speaker: From Policy to Practice: Operationalizing EU AI Act Article 17 Quality Management System in Regulated Industries
09:40 - 09:55
Networking Zone
Ask a Speaker: 927 odpowiedzi, dwie fale, rok różnicy: jak mierzyliśmy adopcję AI w polskim software house i co z tego wynika
09:40 - 09:55
Networking Zone
Ask a Speaker: Agentic Memory at Scale: a practical guide
09:40 - 09:55
Networking Zone
Ask a Speaker: [Special Guest] Fireside Chat
09:40 - 09:55
Networking Zone
Ask a Speaker: [TRACK KEYNOTE] BI isn't dead. It melted.
09:40 - 09:55
Networking Zone
Ask a Speaker: Can agentic AI bridge the gap between Earth Observation and emergency response?
09:40 - 09:55
Networking Zone
Ask a Speaker: Uncovering Clinical AI Deployment Risks with Generative Interventional Attributions
09:40 - 09:55
Networking Zone
Ask a Speaker: GraphRAG in Practice: Five Design Patterns for Reliable Knowledge Extraction
10:25 - 10:40
Networking Zone
Ask a Speaker: [Special Guest] Jak zbudowaliśmy zespół data science, który nie obiecuje, tylko dowozi?
10:25 - 10:40
Networking Zone
Ask a Speaker: Building Nation-Wide Device Identity: The Algorithms Behind Banking Fingerprinting
10:25 - 10:40
Networking Zone
Ask a Speaker: Retrofitting Agentic RAG: The Missing NER Layer for Deterministic Recall. Finding and matching products across domains and millions of pages a week
10:25 - 10:40
Networking Zone
Ask a Speaker: Koniec dashboardów? Jak w 30 minut zacząć zadawać pytania własnym danym
10:25 - 10:40
Networking Zone
Ask a Speaker: Dostawa jutro czy za dwa dni? Geolokalizacyjny ML w czasie rzeczywistym przy 100 000 RPS na Allegro
10:25 - 10:40
Networking Zone
Ask a Speaker: Pół miliarda decyzji cenowych dziennie - produkcyjny ML w optymalizacji aukcji reklamowych
10:25 - 10:40
Networking Zone
Ask a Speaker: Data science methods that drive localization
10:25 - 10:40
Networking Zone
Ask a Speaker: Monitorowanie postępu technologicznego - jak zaplanować migrację na PQC zgodnie z RODO.
10:25 - 10:40
Networking Zone
Ask a Speaker: When AI Agents Enter the Physical World: Humanoids, Robots and the Rise of Dark Factories
10:25 - 10:40
Networking Zone
Ask a Speaker: Open format isn't enough: Building the interoperable data stack
11:15 - 11:30
Networking Zone
Ask a Speaker: Od dokumentu do danych - ekstrakcja informacji finansowych ze sprawozdań z KRS z użyciem AI
11:15 - 11:30
Networking Zone
Ask a Speaker: [PANEL EKSPERCKI] Jak używać danych medycznych? Rola RCMC w ekosystemie zdrowia
11:15 - 11:30
Networking Zone
Ask a Speaker: Demokratyzacja obserwacji Ziemi: Wolny GIS i otwarte dane satelitarne w społecznej analizie pożarów
11:15 - 11:30
Networking Zone
Ask a Speaker: [Track Keynote] Powering the Map of the Law: Distilling LLMs for Scalable Legal Act Processing
11:15 - 11:30
Networking Zone
Ask a Speaker: Same Data & AI Roadmap, Different Stage: How LOT Airlines, Colonnade Insurance and Haddad Brands Modernized Their Data Platforms
11:15 - 11:30
Networking Zone
Ask a Speaker: AI on device / AI beyond the cloud
11:15 - 11:30
Networking Zone
Ask a Speaker: AI w organizacjach
11:15 - 11:30
Networking Zone
Ask a Speaker: Wielowarstwowy system antyfraudowy
11:15 - 11:30
Networking Zone
Ask a Speaker: RAG to za mało: jak mierzyć jakość danych, zanim asystent AI zarekomenduje decyzję biznesową
11:15 - 11:30
Networking Zone
Ask a Speaker: AI works exactly as it should. That's the problem. Why AI adoption exposes flaws that have remained hidden in data for years
11:15 - 11:30
Networking Zone
Ask a Speaker: Silicon Lifeforms. Humans in the loop but who is in control?
11:15 - 11:30
Networking Zone
Ask a Speaker: Let’s Get Ready to Rumble: Power BI + Copilot vs. Databricks AI/BI + Genie
12:00 - 12:15
Networking Zone
Ask a Speaker: Prune the experts you don't serve: targeted shrinking of Mixture-of-Experts LMs
12:00 - 12:15
Networking Zone
Ask a Speaker: The Rise of the "LLM Data Engineer": Preparing Data for Generative AI at Scale
12:00 - 12:15
Networking Zone
Ask a Speaker: When Simpler Models Win: Building a Practical Categorisation & Product-Matching Pipeline
12:00 - 12:15
Networking Zone
Ask a Speaker: [Track Keynote] When the 95% Accuracy Is Not Enough: Explaining, Verifying, and Validating AI Models for Space Applications
12:00 - 12:15
Networking Zone
Ask a Speaker: Can Genie Code Productionize an AI Agent? A Controlled Engineering Experiment
12:00 - 12:15
Networking Zone
Ask a Speaker: Why should Data Governance be AI's best friend
12:00 - 12:15
Networking Zone
Ask a Speaker: Holistyczna analiza klienta pod kątem AML z wykorzystaniem narzędzi AI
12:00 - 12:15
Networking Zone
Ask a Speaker: MolPILE - large-scale, diverse dataset for molecular representation learning
12:00 - 12:15
Networking Zone
Ask a Speaker: Make The Difference: AI ROI
12:00 - 12:15
Networking Zone
Ask a Speaker: AI at scale in healthcare: how to protect models and data against attacks
12:00 - 12:15
Networking Zone
Ask a Speaker: [SPECIAL GUEST] Power BI + MCP: niech agent zrobi model za Ciebie
12:00 - 12:15
Networking Zone
Ask a Speaker: From Legacy to AI-Ready: Spec-Driven Data Modernization with EPAM
12:50 - 13:05
Networking Zone
Ask a Speaker: Semantic Analytics of AI Interactions
12:50 - 13:05
Networking Zone
Ask a Speaker: Wszystko, czego baliście się zapytać o OpenStreetMap, a moze się przydać w pracy.
12:50 - 13:05
Networking Zone
12:50 - 13:05
Networking Zone
Ask a Speaker: From AI North Star Strategy to Deployed AI Agents in Comarch Products
12:50 - 13:05
Networking Zone
Ask a Speaker: Beyond the Ranking: Investigating the Gene Behind a GWAS Association
12:50 - 13:05
Networking Zone
Ask a Speaker: Lokalne modele - LLMy - wrażenia, takeaways z wystawienia lokalnych modeli
12:50 - 13:05
Networking Zone
Ask a Speaker: The Messy Middle of Industrial Data: From Production Systems to Business Insight
12:50 - 13:05
Networking Zone
Ask a Speaker: Czy agenci będą oglądać dashboardy za nas? Dwugłos o roli AI w BI.
12:50 - 13:05
Networking Zone
Ask a Speaker: What Agentic AI taught us about Implementing Agentic AI Solutions
12:50 - 13:05
Networking Zone
Ask a Speaker: One Config-Millions of Events. Building a proprietary solution to configure, deploy, and monitor large-scale experiments
12:50 - 13:05
Networking Zone
Ask a Speaker: Everyone Wanted LLMs. Production Wanted FastText: A Pragmatic Approach to Million-Scale Categorization
12:50 - 13:05
Networking Zone
Ask a Speaker: Jak wycisnąć maksimum z dostępnej infrastruktury? Od optymalizacji agenta głosowego do inferencji na GPU
13:35 - 13:50
Networking Zone
Ask a Speaker: Sequence-Based or Tabular models for Next Best Action Engines? Lessons Learned from the PwC Hackaton
13:35 - 13:50
Networking Zone
Ask a Speaker: Self-Service at Scale: The Data Platform Powering Netflix's ML Feature Pipelines
13:35 - 13:50
Networking Zone
Ask a Speaker: Modality Fusion in Drug Development
13:35 - 13:50
Networking Zone
Ask a Speaker: Elevating conversational AI with domain expertise via Chemical Expert Model Context Protocol (MCP) – Merck KGaA case study
13:35 - 13:50
Networking Zone
Ask a Speaker: [Track Keynote] Bielik v3.1. Prezentacja i omówienie powstawania modelu językowego
13:35 - 13:50
Networking Zone
Ask a Speaker: Neural challenger model for investment risk profiling
13:35 - 13:50
Networking Zone
Ask a Speaker: How will quantum computers change the way we work with AI and Machine Learning?
13:35 - 13:50
Networking Zone
Ask a Speaker: [Special Guest] [Panel]
13:35 - 13:50
Networking Zone
Ask a Speaker: When Gen AI takes advantage over traditional techniques - and when it does
13:35 - 13:50
Networking Zone
Ask a Speaker: Quo vaditis, automata: Autonomous research and coding agents
13:35 - 13:50
Networking Zone
Ask a Speaker: Skalowanie AI w dużych organizacjach - porównanie perspektyw
13:35 - 13:50
Networking Zone
Ask a Speaker: Przyszłość Data & Analytics
15:00 - 15:15
Networking Zone
Ask a Speaker: Using Local Language Models to Assess the Quality of Narrative Data in Operational Risk
15:00 - 15:15
Networking Zone
Ask a Speaker: Od obietnic AI do realnego ROI: Jak sprzedać analitykę konwersacyjną dyrektorowi finansowemu
15:00 - 15:15
Networking Zone
Ask a Speaker: [PANEL] AI-assisted data science: implications for team structure, role descriptions and everyday work
15:00 - 15:15
Networking Zone
Ask a Speaker: Everyone wants more AI. But who is going to govern it?
15:00 - 15:15
Networking Zone
Ask a Speaker: [Special Guest] Evaluation and red teaming of models and agents
15:00 - 15:15
Networking Zone
Ask a Speaker: [Special Guest] How to Break an AI Agent? - Red Teaming Agentic AI Applications.
15:00 - 15:15
Networking Zone
Ask a Speaker: Blank Canvas, Live Data: Painting Dashboards with Snowflake CoCo Agent
15:00 - 15:15
Networking Zone
Ask a Speaker: We shipped a Regulatory Intelligence Deep Agent to 2,500 users. Here's What ML engineering doesn't prepare you for.
15:00 - 15:15
Networking Zone
Ask a Speaker: [PANEL] Czy potrafimy dobrze oceniać AI?
15:00 - 15:15
Networking Zone
Ask a Speaker: Od dojrzałego BI do dojrzałego AI - jak firma z uporządkowaną analityką robi kolejny krok
15:00 - 15:15
Networking Zone
Ask a Speaker: Implementing Data Hubs in Large Organizations: Lessons from the Field
15:00 - 15:15
Networking Zone
Ask a Speaker: Learning to Rank Promotional Artwork with a Vision-Language Model
15:45 - 16:00
Networking Zone
Ask a Speaker: Ludzie i agenty zmyślają - jak pomóc agentom? Pentesterzy i detektywi AI testują hipotezy.
15:45 - 16:00
Networking Zone
Ask a Speaker: One user vector to rule them all - how we put ~200 behavioural features online
15:45 - 16:00
Networking Zone
Ask a Speaker: AI in BI to boost UA
15:45 - 16:00
Networking Zone
Ask a Speaker: Embedding physics priors into machine learning: lessons on modeling process dynamics from a steam-saving project in food manufacturing.
15:45 - 16:00
Networking Zone
Ask a Speaker: When AI Meets Compliance: Building an LLM Agent for Financial KYC Process
15:45 - 16:00
Networking Zone
Ask a Speaker: Kto ma czerwony przycisk? O kontroli, autonomii AI na produkcji
15:45 - 16:00
Networking Zone
Ask a Speaker: Hearing diagnosis in 1.5 seconds: wavelet features and classical ML on 5,652 clinical ABR signals
15:45 - 16:00
Networking Zone
Ask a Speaker: Beyond the model: Designing AI systems that solve real problems. Three practical lessons and what they mean for the evolving role of the data scientist
15:45 - 16:00
Networking Zone
Ask a Speaker: How to Evaluate AI When There Is No Correct Answer: 3 CVs, 600 Interviews, and One Hidden Prompt
15:45 - 16:00
Networking Zone
Ask a Speaker: [TOPIC SOON: SOFTSERVE]
15:45 - 16:00
Networking Zone
15:45 - 16:00
Networking Zone
Ask a Speaker: Optimizing Ultralytics YOLO26: From 41 FPS up to 922 FPS in 30 minutes
16:35 - 16:50
Networking Zone
Ask a Speaker: Snowflake CoCo — jak AI agent zmienia codzienną pracę data engineera
16:35 - 16:50
Networking Zone
Ask a Speaker: Scaling Enterprise AI
16:35 - 16:50
Networking Zone
Ask a Speaker: Gdzie jestem, gdzie idę? Rola analityka BI w organizacji
16:35 - 16:50
Networking Zone
Ask a Speaker: Odporni na hype: Dlaczego w optymalizacji przestrzeni Empiku postawiliśmy na klasyczną ekonometrię i modele mieszane?
16:35 - 16:50
Networking Zone
Ask a Speaker: Czy PowerPoint już umarł?
16:35 - 16:50
Networking Zone
Ask a Speaker: Embeddings for Earth: How Foundation Models Are Reshaping GeoAI
16:35 - 16:50
Networking Zone
Ask a Speaker: Don't Make the Network Rediscover Anatomy: Computer Vision for Very Early-Stage Lung Cancer in LDCT
16:35 - 16:50
Networking Zone
Ask a Speaker: Od dokumentów do wiedzy. Kulisy budowy bazy wiedzy AI w dużej organizacji
16:35 - 16:50
Networking Zone
Ask a Speaker: 22 Hits to the Right Side of the Head: Turning Raw Impact Telemetry into Decisions Combat Athletes Trust; Without Overclaiming
16:35 - 16:50
Networking Zone
Ask a Speaker: Six safety gates across 13 stages: the latency and review cost of a governed content generation pipeline
16:35 - 16:50
Networking Zone
Ask a Speaker: [Special Guest] Writing in a Genre Is a Specialised Task: LLM Domain and Task Adaptation for Polish Official Writing
16:35 - 16:50
Networking Zone
Ask a Speaker: AI Agents Meet Predictive Modeling: Didn't See That Coming
17:20 - 17:35
Networking Zone
Ask a Speaker: Automating the process of scientific discovery in practice: a deep dive into Molecule.one and OpenAI near-autonomous discovery in organic chemistry
17:20 - 17:35
Networking Zone
Ask a Speaker: What Happens When Time Series Foundation Models Leave the Benchmark and Enter the Real World
17:20 - 17:35
Networking Zone
Ask a Speaker: Try. Fail. Codify. Repeat. — 100 Days of AI as an Engineering Manager
17:20 - 17:35
Networking Zone
Ask a Speaker: Agentic AI w sektorze finansowym
17:20 - 17:35
Networking Zone
Ask a Speaker: The Hidden Failure Point in AI Projects: Externalizing Tacit Knowledge Before Scaling from PoC to Production
17:20 - 17:35
Networking Zone
Ask a Speaker: From Assistants to Executable Workflows: A New Paradigm for Industrial AI
17:20 - 17:35
Networking Zone
Ask a Speaker: Autonomous Enterprise: Reference Architecture for Multi-Agent AI Platforms
17:20 - 17:35
Networking Zone
Ask a Speaker: [Track Keynote] Scaling the Brain: ML-Driven Vertical Autoscaling for GKE Control Plane
17:20 - 17:35
Networking Zone
Ask a Speaker: FertiPredicta: predykcja skuteczności IVF przed i w kolejnych cyklach leczenia
17:20 - 17:35
Networking Zone
Ask a Speaker: Where the RAM Goes: Sizing a Billion-Vector Index from 3.3 TB Down to 65 GB
17:20 - 17:35
Networking Zone
Ask a Speaker: Prawdziwa podróż klienta wygląda inaczej, niż myślisz. Jak wykorzystujemy process mining do analizy ścieżek klienta w kontaktach z bankiem
17:20 - 17:35
Networking Zone
Ask a Speaker: [Special Guest] Beyond the Model: Practical Considerations for Deploying AI-Driven Robots in the Real World
18:05 - 18:20
Networking Zone
Ask a Speaker: Czy no/low-code automatyzacja jaką znamy, jeszcze ma sens w czasach AI?
18:05 - 18:20
Networking Zone
Ask a Speaker: Myśleć, mówiąc: trójstrumieniowa architektura dla niskolatencyjnych agentów głosowych.
18:05 - 18:20
Networking Zone
Ask a Speaker: AI z lotu ptaka
18:05 - 18:20
Networking Zone
Ask a Speaker: Beyond Model Accuracy: Turning Healthcare Infection Risk Prediction into Real Process Improvement
18:05 - 18:20
Networking Zone
Ask a Speaker: Zero-ETL czy tylko mniej ETL? Mirroring, Copy Job i Pipelines w Microsoft Fabric
18:05 - 18:20
Networking Zone
Ask a Speaker: FinAIOps, using AI to predict, explain and optimise cloud spent
18:05 - 18:20
Networking Zone
Ask a Speaker: ISO 9001 without a consultancy: how AI agents built – and now run – a certified quality management system
18:05 - 18:20
Networking Zone
Ask a Speaker: Representation learning approach to taxonomy analysis
18:05 - 18:20
Networking Zone
Ask a Speaker: Energy storage – estimating investment profitability
18:05 - 18:20
Networking Zone
Ask a Speaker: Where Did This Value Come From? Governing AI-Assisted Data Enrichment in Systems of Record
18:05 - 18:20
Networking Zone
Ask a Speaker: Kiedy atrybucja kłamie: jak przebudowaliśmy pomiar efektywności marketingu z wykorzystaniem BigQuery i Marketing Mix Modeling
19:00 - 00:00
Kandela Bar (Elektrownia Powiśle)
19:00 - 21:00
Kandela Bar (Elektrownia Powiśle)

We created an attractive agenda for everyone - professionals and experts as well as decision makers

An event that should be one of the main points on the calendar of anyone associated with the data area

The Data Science Summit explores all current and important threads, such as the security of AI systems, the explainability and interpretability of predictive models, aspects relevant during the research phase as well as the implementation phase

The conference addresses the use of data in the organization in an attractive way for executives - from project manager to board member

DSS is not only interesting speeches, but also inspiring talks, exchange of experiences and a great opportunity to network with fantastic people

All those involved in the development, implementation and maintenance of data solutions will find a lot of interest here

DSS over the past few years has become the largest event of its kind in the CEE region. Here you will meet industry leaders and thousands of others from the *data* community.

It's great that the agenda is divided into as many as 3 formats - online live, stationary and as a VoD library. It's best to be on-site but regardless of how and when you prefer to watch, you're sure to learn a lot here

Online
Thursday, 26.11.2026
available for logged-in users.
Onsite
Friday, 27.11.2026
Conference Centre
al. Ks. Józefa Poniatowskiego 1
03-901 Warsaw
entrance through gate no. 1 from Zieleniecka Avenue


































Main Sponsor
Special Sponsor
Advanced Technology Sponsor
Gold Sponsor
Silver Sponsors
Innovation Partner
At the event you will meet employers and providers of IT / data science products and services who will introduce you to their company profiles and recruitment offers.















Main organizer
Since 2005, we have been supporting the Polish IT / data science community in its development. We are not only the organizer, but also the initiator of many significant events aimed at industry professionals and students, such as Warsaw IT Days, Data Science Summit, The Hack Summit, Ya!vaConf, MS Tech Summit, Warsaw IT Days Expo and many meetups. We also help organize IT and data science conferences for companies / organizations interested in such activities.
Data Science Warsaw is the largest data science community in Poland, based in Warsaw. We are a non-profit organization working on the exchange of ideas and knowledge about data science, data engineering and so-called artificial intelligence. Our goal is to discuss tools, technology and business opportunities related to data collection, processing and visualization, as well as machine learning and deep learning.













