Python Summit

19.10.2026

DevAI: Main Stage (Online)

Program Governance:

Chairmans:

08:30 - 09:00

Online Live

Difficulty:

Official Opening

dr inż. Michał Bijata - Prezes zarządu, Fundacja Academic Partners

dr. hab. Dominik Batorski - Przewodniczący Rady Programowej, DevAI by DSS

Rafał Małanij - Data Manager, CBRE Polska

Rafał Korzeniewski - Software Engineer, Python Trainer, PyWaw Coorganizer, ITMAGINATION

09:00 - 09:35

Online Live

Difficulty:

Replacing the LLM in Your Agent with a Small Language Model: A Practical Guide

Jacek Golebiowski - CTO & Co-founder, distil labs

small language models, agents, LLM, fine-tuning, model distillation

09:40 - 10:15

Online Live

Difficulty:

Frozen at Training Time: Building Production Memory Layers for Stateful AI Agents

Andrey Goryainov - Founding Engineer, Olodin

AI Agents, Agent Memory, Continual Learning, Context Engineering, Machine Learning Engineering, Enterprise AI Architecture

10:20 - 10:55

Online Live

Difficulty:

Beyond Agents: A New Architecture for Reproducible Code Generation Workflows

Catalin Tudose - Java and Web Technologies Expert, Luxoft

Vladimir Sonkin

Data Modeling, AI Ontology, Agentic Workflows, Entity Resolution, Semantic Layer, Enterprise AI

11:00 - 11:35

Online Live

Difficulty:

How to Keep Coding Agents from Breaking the Repository: Permission Boundaries, Verification Loops, and Human Control

Andrew Kozlov - Principal Agentic Engineer, ENDGAME

coding agents, Claude Agent SDK, agent engineering, tool permissions, verification loops, human-in-the-loop, repository safety, developer tools

11:40 - 12:15

Online Live

Difficulty:

Deciding What Is "Correct": Verification as the New Role of the Engineer in the Agentic Era

Dmitrii Nikitin - Android Team Leader, Quadcode

AI Agents, Software development, AI adoption

12:20 - 12:55

Online Live

Difficulty:

Agentic AI in Web3 Security: Autonomous Defenders and the Risks They Introduce

Aleksander Kryukov - Security Engineer, Hypernative

Agentic AI, Web3 Security, Smart Contract Auditing, On-chain Incident Response, Adversarial Machine Learning, Threat Detection

12:55 - 13:55

Online Live

Difficulty:

Lunch break

13:55 - 14:30

Online Live

Difficulty:

Agents That Never Stop: The Pattern Behind Self-Improving Systems

Nick Malozemov - Founding Engineer, Olodin

AI Agents, Autonomous Loops, System Architecture, Automation, MLOps, System Robustness

14:35 - 15:10

Online Live

Difficulty:

Wirtualni pracownicy w SDLC: jak delegować pracę agentom AI bez utraty kontroli

Mateusz Ciuła - Founder & CEO, Japonics

agenci AI, agent engineering, SDLC, automatyzacja, ClickUp, context engineering, plan-then-execute, code review, halucynacje AI, human-in-the-loop, audit trail, delegacja zadań

15:15 - 15:50

Online Live

Difficulty:

Vibes Don't Scale

Georgios Diamantopoulos - Founder, Zero to MVP

vibe coding, software engineering

15:55 - 16:30

Online Live

Difficulty:

Co naprawdę psuje się w agentach AI na produkcji: pułapki decyzyjne, pozorna racjonalność i cichy dryf modeli

Aleksander Fafuła - AI Architect, funduszAI.pl

Multi-agent systems, LLM, Agent Engineering, Behavioral Finance, AI Ops, Financial Data Mining, Autonomous Agents

16:35 - 17:10

Online Live

Difficulty:

Więcej niż Chat: Jak modele LLM rewolucjonizują inżynierię cech w tradycyjnych systemach Machine Learning

Aneta Maksymiuk - Principal Analytical Consultant, SAS Institute

Piotr Skarzyński

LLM, feature engineering

17:15 - 17:50

Online Live

Difficulty:

MicroAgents: Applying Microservices Architecture to Production AI Agents

Ashish Shubham - VP Engineering, ThoughtSpot Inc

AI agents, MicroAgents, agent infrastructure, multi-agent orchestration, MCP

20.10.2026

AI in Software Development (Onsite)

Program Governance:

Chairmans:

08:00 - 08:30

Centrum Nauki Kopernik

Difficulty:

Otwarcie rejestracji | Registration opening

08:30 - 09:00

Room 1

Difficulty:

Official Opening

09:00 - 09:35

Room 1

[KEYNOTE] When System Knowledge Disappears: How AI Helped Us Regain Control of a Project

Paweł Ekk-Cierniakowski - Senior Manager Data Science | Trainer, SoftwareOne | Sages

knowledge recovery, legacy systems, GitHub Copilot, system analysis, Human-AI collaboration, reverse engineering

09:45 - 10:20

Room 1

Difficulty:

How AI Native development is reshaping the team structure

Bartosz Fijałkowski - AI Native Engineering Lead & Head of .NET Development, GFT Technologies

AI, AINative, Leadership, Teams, SDD, SDLC

10:35 - 11:10

Room 1

Difficulty:

Non deterministic tools applied to non deterministic process - code review harness @ Saily

Piotr Szadkowski - Senior iOS Developer, Saily, powered by Nord Security

Marcin Jakuszko - , Saily, powered by Nord Security

11:20 - 11:55

Room 1

Difficulty:

Agenci AI na straży faktów - generatywna AI w analizie danych i walce z dezinformacją

Tomasz Makowski - Specjalista ds. Machine Learning, Asseco

12:10 - 12:45

Room 1

Difficulty:

Building software with AI in 2026: a practical playbook for protecting and monetizing what you ship

Marcin Ręgorowicz - Attorney-at-Law / Co-founder, N77 LEGAL

AI-assisted software, AI-assisted coding, software ownership, dataset licenses, model inputs, code provenance, dataset provenance, open source, software composition analysis, RAG, fine-tuning, copyright, intellectual property, trade secrets, software monetization

12:55 - 13:30

Room 1

Difficulty:

Beyond the repository - agentic portfolio management

Tomasz Guściora - DemystifAI, CEO

agentic ai, scaling, ai-dev, codex, claude code, ohmypi, pi, agents, LLM, AI-assisted software development, software development, CC, AI SDLC, SDLC

13:30 - 14:20

Centrum Nauki Kopernik

Difficulty:

Lunch break

14:20 - 14:55

Room 1

Difficulty:

Agentic coding w prawdziwym kodzie: gdy dostarczanie przestaje być wąskim gardłem

Mateusz Gostański - Staff Software Engineer, Alokai

agentic coding, claude code, ai coding

15:05 - 15:40

Room 1

Difficulty:

We Have Met the Enemy and It Is Us: How AI Productivity Outpaces System Reliability

Borys Stokalski - Expert, GPAI/OECD.AI, RETHINK

responsible AI, AI engineering, generative development

EN HOT

15:55 - 16:30

Room 1

Difficulty:

What Coding Agents Still Get Wrong

Michał Prządka - Independent AI Consultant, Michał Prządka

coding agents, AI-assisted development, Claude Code, Codex, OpenCode, LLM evaluation, agent failure modes, developer productivity, effort estimation, AI reliability

16:40 - 17:15

Room 1

Difficulty:

How a "Don't Touch It" Legacy App Became AI-First

Michał Pobuta - Software Engineer, Heineken

legacy code, AI-First, refactoring agents, AI-assisted coding, dev workflow, code graph, markdown documentation, codebase analysis, AI agents, technical debt, .NET MAUI, mobile development

EN HOT

17:25 - 18:00

Room 1

Difficulty:

From Search to Booking: Building Trusted Android Workflows with App Functions

Andrii Veremiienko - Android Software Engineer, Revolut

Android App Functions, agentic Android, AI assistants, accessibility, accessible travel, deaf travellers, Kotlin, typed app functions, domain logic, authentication, booking flows, partial failures, ambiguous requests, availability management, native UI handoff, business rules, conversational interfaces, AI integration, Android architecture, user safety, accessible UX, production Android, travel technology

EN HOT

AI Systems in Production (Onsite)

Program Governance:

Chairmans:

09:00 - 09:35

Room 2

Difficulty:

Gdzie kończy się agent a zaczyna kod — niezawodność systemów multiagentowych na produkcji.

Agata Hanas-Szadkowska - Senior AI Scientist, Pearson

systemy multiagentowe, agenty LLM, niezawodność, orkiestracja agentów

No seats in selected package. Seats available in other packages.

09:45 - 10:20

Room 2

Difficulty:

Last Year's Frontier AI Now Fits on Your Laptop

Piotr Migdał - Founding Engineer, Quesma

local models, qwen, apple silicon, llama.cpp, quantization, performance, benchmark

EN HOT Last places in basic packages

10:35 - 11:10

Room 2

Difficulty:

[KEYNOTE] Retrofitting Agentic RAG: The Missing NER Layer for Deterministic Recall. Finding and matching products across domains and millions of pages a week

Jan Majewski - Lead AI Engineer, Minerva Technology

NER, RAG, Product Search, Context Engineering

EN HOT Last places in basic packages

11:20 - 11:55

Room 2

Difficulty:

Vibe-Coding a Second Brain: What an AI Agent Actually Shipped vs the Demo

Yury Vashchylau - Founder, yuryv.info

agentic coding, AI coding agents, AI-assisted development, Kotlin, Android, Jetpack Glance widgets, LLM integration, developer productivity, code review guardrails, verification, note capture, failure modes

EN No seats in selected package. Seats available in other packages.

12:10 - 12:45

Room 2

Difficulty:

How many agents can one GPU actually serve? Measuring the multi-user context wall before it hits production

Marek K. Zielinski - Co-founder, Ryliq

LLM serving, multi-user, concurrency, KV cache, context length, vLLM, SGLang, llama.cpp, throughput, TTFT, speculative decoding, MoE, agentic, GPU benchmarking, capacity planning

EN Last places in basic packages

12:55 - 13:30

Room 2

Difficulty:

Talk With Your Data: Building AI-Powered Reporting

Alper Ebicoglu - Co-founder, Volosoft

AI, Reporting, SQL, BI, Data Visualization

EN Last places in basic packages

13:30 - 14:20

Centrum Nauki Kopernik

Difficulty:

Lunch break

14:20 - 14:55

Room 2

Difficulty:

The Extinct Attractor: What Happened to "I Don't Know" in Instruction-Tuned LLMs

Piotr Gieszcz - Data Scientist, AI Architect, atlab.waw.pl

LLMs, Hallucination detection, Abstention state, Instruction tuning, Autoregressive generation dynamics

EN Last places in basic packages

15:05 - 15:40

Room 2

Difficulty:

From Trace to Fix: Debugging Production AI Agents with Genie Code and MLflow

Maciej Rubczyński - AI Engineer, Entrada AI

Genie Code, MLflow, Databricks, AI agents, AgentOps, tracing, evaluation, observability, LangGraph, Genie Agents, Unity Catalog, production reliability, debugging

EN Last places in basic packages

15:55 - 16:30

Room 2

Difficulty:

Symbioza Predictive ML z LangGraph: wzorce architektoniczne, eksperci i pułapki ewaluacji

Tomasz Bąk - CTO, AIForecast

LangGraph, Multi-Agent Systems, Time-Series ML, Tool Calling, MLflow, LLMOps, Financial Engineering

Last places in basic packages

16:40 - 17:15

Room 2

Difficulty:

Od alertów do diagnozy: architektura autonomicznego agenta SRE

Ryszard Szwajlik - AI Technical Lead, j-labs

Jakub Gajski - Lider Techniczny AI, j-labs

AI Agents, Site Reliability Engineering, SRE, DevOps, incident diagnosis, Microsoft Agent Framework, Magentic orchestration, Magentic-One, multi-agent systems, ITBench, observability, context management, AI safety, auditability

Last places in basic packages

17:25 - 18:00

Room 2

Difficulty:

When Data can Act: Security Data Risks in LLM Systems.

Jan Broniowski - AI Security Consultant, SafeDescent

AI Security, Data Risks, RAG, threat modeling

EN No seats in selected package. Seats available in other packages.

AI Agents (Onsite)

Program Governance:

Chairmans:

13:30 - 14:20

Centrum Nauki Kopernik

Difficulty:

Lunch break

14:20 - 14:55

Room 5

Difficulty:

[KEYNOTE] Text-to-Diagram for AI Systems: Designing a DSL an LLM Gets Right

Oleg Żero - freelance consultant, Zerowithdot Oleg Żero

AI solution architecture, architecture notation, text-to-diagram, diagram-as-code, LLM structured output, agent engineering

EN Last places in basic packages

15:05 - 15:40

Room 5

Difficulty:

Intent Before Intelligence: Fine-Tuning Small Models for Smarter Routing

Serg Masis - Principal AI Scientist, Syngenta

fine-tuning, routing, agents, ai

EN Last places in basic packages

15:55 - 16:30

Room 5

Difficulty:

When AI Agents Fail in Production: Tracing, Budgets, and Tool Execution Ledgers

Mikalai Saskavets - GenAI Engineer / Python Developer, Coherent Solutions

AgentOps, AI agents, observability, tracing, tool ledger, idempotency, LLM, production reliability, evaluation, retries, cost budgets, Python, cloud-native systems

EN Last places in basic packages

16:40 - 17:15

Room 5

Difficulty:

Generative UI on Ollama Cloud: forcing a small open-weight model into legal moves

Paweł Twardziak - Senior Software Engineer & AI Engineer, Devs 3.0

small language models, open-weight models, Ollama Cloud, generative UI, structured output, tool calling, function calling, self-correction, ground-truth validation, LangGraph, CopilotKit, agentic UI

EN Last places in basic packages

17:25 - 18:00

Room 5

Difficulty:

AI end-to-end claim handling: od zgłoszenia szkody do błyskawicznej decyzji

Marta Łukjanowicz - Product Owner, ERGO Hestia

Piotr Brodziak - Software Development Team Lead, Ergo Hestia

AI agents, document processing, claims automation, human in the loop, process automation

Last places in basic packages

Roundtables (Onsite)

10:35 - 11:10

Room RT

Difficulty:

Autonomous Research w softdevie

Jakub Nowacki - Engineering Manager and Technical Leader in AI/ML, Google

No seats in selected package. Seats available in other packages.

11:20 - 11:55

Room RT

Difficulty:

Co agenci AI muszą wiedzieć o naszych systemach - i skąd mają czerpać tę wiedzę?

Andrzej Sobczak - Kierownik Zakładu Zarządzania IT, Szkoła Główna Handlowa w Warszawie

No seats in selected package. Seats available in other packages.

12:10 - 12:45

Room RT

Difficulty:

How to plan and control AI expenses

Paweł Ekk-Cierniakowski - Senior Manager Data Science | Trainer, SoftwareOne | Sages

No seats in selected package. Seats available in other packages.

12:55 - 13:30

Room RT

Difficulty:

How to use AI for code & design without generating slop?

Piotr Migdał - Founding Engineer, Quesma

No seats in selected package. Seats available in other packages.

14:20 - 14:55

Room RT

Difficulty:

AI Agents - technicznie

Łukasz Czarnecki

No seats in selected package. Seats available in other packages.

15:05 - 15:40

Room RT

Difficulty:

Kill Switch for AI: bare minimum or just the start?

Wojciech Gawroński - Developer Advocate, Unleash

Last places in basic packages

Ask a speaker (Onsite)

09:40 - 09:55

Networking Zone

Difficulty:

Ask a Speaker: [KEYNOTE] When System Knowledge Disappears: How AI Helped Us Regain Control of a Project

Paweł Ekk-Cierniakowski - Senior Manager Data Science | Trainer, SoftwareOne | Sages

09:40 - 09:55

Networking Zone

Difficulty:

Ask a Speaker: Gdzie kończy się agent a zaczyna kod - niezawodność systemów multiagentowych na produkcji.

Agata Hanas-Szadkowska - Senior AI Scientist, Pearson

10:25 - 10:40

Networking Zone

Difficulty:

Ask a Speaker: Scaling Agentic Workflows: How We Built an MCP-Backed Engineering Operating Model

Marat Kenzhebulatov - Senior Software Engineer, Booking.com

10:25 - 10:40

Networking Zone

Difficulty:

Ask a Speaker: Last Year's Frontier AI Now Fits on Your Laptop

Piotr Migdał - Founding Engineer, Quesma

11:15 - 11:30

Networking Zone

Difficulty:

Ask a Speaker: Autonomous Research w softdevie

Jakub Nowacki - Engineering Manager and Technical Leader in AI/ML, Google

11:15 - 11:30

Networking Zone

Difficulty:

Ask a Speaker: Non deterministic tools applied to non deterministic process - code review harness @ Saily

Piotr Szadkowski - Senior iOS Developer, Saily, powered by Nord Security

Marcin Jakuszko - , Saily, powered by Nord Security

11:15 - 11:30

Networking Zone

Difficulty:

Ask a Speaker: [KEYNOTE] Retrofitting Agentic RAG: The Missing NER Layer for Deterministic Recall. Finding and matching products across domains and millions of pages a week

Jan Majewski - Lead AI Engineer, Minerva Technology

12:00 - 12:15

Networking Zone

Difficulty:

Ask a Speaker: Vibe-Coding a Second Brain: What an AI Agent Actually Shipped vs the Demo

Yury Vashchylau - Founder, yuryv.info

12:00 - 12:15

Networking Zone

Difficulty:

Ask a Speaker: [TOPIC SOON]

12:00 - 12:15

Networking Zone

Difficulty:

Ask a Speaker: [TOPIC SOON]

12:50 - 13:05

Networking Zone

Difficulty:

Ask a Speaker: Building software with AI in 2026: a practical playbook for protecting and monetizing what you ship

Marcin Ręgorowicz - Attorney-at-Law / Co-founder, N77 LEGAL

12:50 - 13:05

Networking Zone

Difficulty:

Ask a Speaker: How many agents can one GPU actually serve? Measuring the multi-user context wall before it hits production

Marek K. Zielinski - Co-founder, Ryliq

12:50 - 13:05

Networking Zone

Difficulty:

Ask a Speaker: [TOPIC SOON]

Paweł Ekk-Cierniakowski - Senior Manager Data Science | Trainer, SoftwareOne | Sages

13:35 - 13:50

Networking Zone

Difficulty:

Ask a Speaker: Talk With Your Data: Building AI-Powered Reporting

Alper Ebicoglu - Co-founder, Volosoft

13:35 - 13:50

Networking Zone

Difficulty:

Ask a Speaker: From Jobs to Workflows: Architecting the Human Layer of AI-Assisted Tech Teams

Vera Solomatina - SVP People & Culture, inDrive

13:35 - 13:50

Networking Zone

Difficulty:

Ask a Speaker: [TOPIC SOON]

15:00 - 15:15

Networking Zone

Difficulty:

Ask a Speaker: AI Agents - technicznie

Łukasz Czarnecki

15:00 - 15:15

Networking Zone

Difficulty:

Ask a Speaker: [KEYNOTE] Text-to-Diagram for AI Systems: Designing a DSL an LLM Gets Right

Oleg Żero - freelance consultant, Zerowithdot Oleg Żero

15:00 - 15:15

Networking Zone

Difficulty:

Ask a Speaker: Agentic coding w prawdziwym kodzie: gdy dostarczanie przestaje być wąskim gardłem

Mateusz Gostański - Staff Software Engineer, Alokai

15:00 - 15:15

Networking Zone

Difficulty:

Ask a Speaker: The Extinct Attractor: What Happened to "I Don't Know" in Instruction-Tuned LLMs

Piotr Gieszcz - Data Scientist, AI Architect, atlab.waw.pl

15:45 - 16:00

Networking Zone

Difficulty:

Ask a Speaker: [TOPIC SOON]

15:45 - 16:00

Networking Zone

Difficulty:

Ask a Speaker: We Have Met the Enemy and It Is Us: How AI Productivity Outpaces System Reliability

Borys Stokalski - Expert, GPAI/OECD.AI, RETHINK

15:45 - 16:00

Networking Zone

Difficulty:

Ask a Speaker: Intent Before Intelligence: Fine-Tuning Small Models for Smarter Routing

Serg Masis - Principal AI Scientist, Syngenta

15:45 - 16:00

Networking Zone

Difficulty:

Ask a Speaker: From Trace to Fix: Debugging Production AI Agents with Genie Code and MLflow

Maciej Rubczyński - AI Engineer, Entrada AI

16:35 - 16:50

Networking Zone

Difficulty:

Ask a Speaker: Symbioza Predictive ML z LangGraph: wzorce architektoniczne, eksperci i pułapki ewaluacji

Tomasz Bąk - CTO, AIForecast

16:35 - 16:50

Networking Zone

Difficulty:

Ask a Speaker: What Coding Agents Still Get Wrong

Michał Prządka - Independent AI Consultant, Michał Prządka

16:35 - 16:50

Networking Zone

Difficulty:

Ask a Speaker: When AI Agents Fail in Production: Tracing, Budgets, and Tool Execution Ledgers

Mikalai Saskavets - GenAI Engineer / Python Developer, Coherent Solutions

17:20 - 17:35

Networking Zone

Difficulty:

Ask a Speaker: Generative UI on Ollama Cloud: forcing a small open-weight model into legal moves

Paweł Twardziak - Senior Software Engineer & AI Engineer, Devs 3.0

17:20 - 17:35

Networking Zone

Difficulty:

Ask a Speaker: How a "Don't Touch It" Legacy App Became AI-First

Michał Pobuta - Software Engineer, Heineken

17:20 - 17:35

Networking Zone

Difficulty:

Ask a Speaker: Od alertów do diagnozy: architektura autonomicznego agenta SRE

Ryszard Szwajlik - AI Technical Lead, j-labs

Jakub Gajski - Lider Techniczny AI, j-labs

18:05 - 18:20

Networking Zone

Difficulty:

Ask a Speaker: AI end-to-end claim handling: od zgłoszenia szkody do błyskawicznej decyzji

Marta Łukjanowicz - Product Owner, ERGO Hestia

Piotr Brodziak - Software Development Team Lead, Ergo Hestia

18:05 - 18:20

Networking Zone

Difficulty:

Ask a Speaker: When Data can Act: Security Data Risks in LLM Systems.

Jan Broniowski - AI Security Consultant, SafeDescent

18:05 - 18:20

Networking Zone

Difficulty:

Ask a Speaker: From Search to Booking: Building Trusted Android Workflows with App Functions

Andrii Veremiienko - Android Software Engineer, Revolut

Afterparty

19:00 - 00:00

Kandela Bar, Elektrownia Powiśle

Difficulty:

Afterparty

19:30 - 20:30

Kandela Bar, Elektrownia Powiśle

Difficulty:

Networking Session

VoD

DevAI 2026 (VoD)

08:00 - 08:10

Online VoD

Difficulty:

Generative UI: These Slides Don't Exist Yet

Georgios Diamantopoulos - Founder, Zero to MVP

generative ui, AI

08:10 - 08:20

Online VoD

Difficulty:

Jak zbudować agentyczny zespół analityczno-architektoniczny

Andrzej Sobczak - Kierownik Zakładu Zarządzania IT, Szkoła Główna Handlowa w Warszawie

System wieloagencki, Wirtualne biuro architektury, Projektowanie architektury, Claude Code

08:20 - 08:30

Online VoD

Difficulty:

Your Agent Has 50 Tools. Now What?

Sofiya Nasiakaila - AI Systems Engineer, Alix

AI agents, tool selection, MCP, tool retrieval, agent engineering, prompt engineering, context engineering, tool calling, LLM evaluation, semantic retrieval, agent architecture, tool discovery, production AI systems

08:30 - 08:40

Online VoD

Difficulty:

Your Pipeline Runs in 4 Minutes. The Business Waited 3 Weeks: Measuring Flow, Not Uptime, in Data & ML Platforms

Ignacio Ceaglio - Senior Lead Data Engineer, Freelancer

data engineering, MLOps, observability, production reliability, flow metrics, flow efficiency, value stream, data pipelines, pipeline reliability, technical debt, data platform, DevOps for data, lead time, bottleneck analysis, stakeholder trust, data observability, project to product, engineering metrics, data quality, business alignment

08:40 - 08:50

Online VoD

Difficulty:

The Judge Problem: Scaling Production Agents with Dynamic Multi-Model Verification

Vsevolod Voropaev - Founding Engineer, Olodin

AI Agents, LLM-as-a-judge, Production Reliability, Multi-Armed Bandits, LLM Routing, Evaluation Architecture

08:50 - 09:00

Online VoD

Difficulty:

AI needs DevOps

Marcin Gastol - Senior DevOps Engineer, Microsoft MVP

DevOps,AI,Bicep,Azure

09:00 - 09:10

Online VoD

Difficulty:

LLMs as Prediction Models: How Prompts Bound the Space of Possible Answers

Dariusz Wawer - Head of Technology and AI, Pretius

AI-assisted coding, context engineering, LLM cost, answer space, prompt boundaries

09:10 - 09:20

Online VoD

Difficulty:

Jak NIE robić projektów w Data Science

Krzysztof Wojewodzic - CEO, Escola S.A.

projekty AI, Data Science, zarządzanie projektami, case study, błędy w projektach AI, AI Act, dane, kompetencje zespołu, lekcje z projektów

09:20 - 09:30

Online VoD

Difficulty:

Skalowalny web scraping w erze LLM: kiedy AI faktycznie pomaga, a kiedy tylko przepala budżet

Kamil Kwapisz - AI Engineer & Consultant, KamilKwapisz.com

agentic ai, web scraping, self-healing, large scale, redukcja kosztów LLM

09:30 - 09:40

Online VoD

Difficulty:

Automating CI/CD Pipeline Remediation: Classifying and Triaging Flaky Tests via LLM Log Parsing

Serhii Yakovenko - Engineering Manager, Toptal

LLM, CI/CD, DevOps, Automation, Log Parsing, Pipeline Reliability

09:40 - 09:50

Online VoD

Difficulty:

AI-Driven Test Automation with Claude Code: Real-World Lessons from Production

Ivan Balashov - Team Lead/ Senior Software Engineer, SoftSwiss

AI Test Automation, Claude Code, Developer Tools, Generative AI, Unit Testing, API Testing, LLM Integration, Developer Productivity

09:50 - 10:00

Online VoD

Difficulty:

AI nie naprawi waszego procesu. Za to zrobi go szybszym

Kamil Banach - Software Engineer, banach.net.pl

AI-assisted development, AI coding agents, Copilot, Claude, Codex, AGENTS.md, code review, developer productivity

10:00 - 10:10

Online VoD

Difficulty:

Scraping Without the Bill: Architecting Resilient Agentic Web Pipelines with Local Language Models

Kate Papis

local LLM, web scraping, agentic pipeline, data extraction, schema validation, self-healing, inference cost, open-weight models, pipeline architecture

10:10 - 10:20

Online VoD

Difficulty:

Harness Engineering - Principles and Practices

Oresztesz Margaritisz - Chief Engineer II, EPAM Systems

AIArchitecture, AIEngineering, GenAI-SDLC, Platform-Engineering, SRE

10:20 - 10:30

Online VoD

Difficulty:

When a Dash Means Zero: Giving AI Trustworthy Context from Open Government Data

Tom Kaltofen - CTO, mloda

RAG, context engineering, open government data, data contracts, CSV parsing, DCAT metadata, property-based testing, declarative data connectors, data quality, mloda, Python, open source