Join us for an exciting Google Dev Fest in Wroclaw at Uniwersytet Ekonomiczny Wroclaw! Over two days, dive into the world of technology with insightful discussion panels covering AI, Web, Mobile, and Cloud. Start your Friday evening with four stimulating discussions, each prefaced by a diverse 10-minute keynote. Then, continue on Saturday with a full day of engaging talks and hands-on workshops. Don't miss this opportunity to connect with like-minded developers and enhance your skills. Mark your calendar for November 13-14, 2026, and become part of the growing tech community now!
AGENDA
Friday, 13 November: Discussion panels & keynotes
16:00 Web panel - moderator: Adrian Romański; panelists: Dominika Zajac, Muhammad Ahsan Ayaz, Jarosław Żołnowski, Julia Rapczyńska, Jakub Andrzejewski
16:50 Keynote - TBA
17:00 Cloud panel - moderator: Michał Jabłoński; panelists: Goran Minov, Çalgan Aygün, Pawel Ostromecki, Piotr Biniecki, Ankur Roy
17:50 Keynote - TBA
18:00 Mobile panel - moderator: Artur Skrzypczyk; panelists: Piotr Prus, Elaine Dias Batista, Dominik Šimoník, Oleksandr Tilnyi, Renata Wietrzyńska
18:50 Keynote - TBA
19:00 AI panel - moderator: Dawid Perdek; panelists: Greg Radzio, Łukasz Jaźwa, Paweł Kubiak, Marcin Czarkowski, Tarun Jain
19:50 Keynote - TBA
Saturday, 14 November: Workshops
7:30 Welcome
8:00-8:45 Cloud - Goran Minov - Zero Infrastructure, Zero Phishing: Building a Serverless Security Framework on GCP
9:00-9:45 Robotics - Łukasz Janiec - From Hallucination to Proof: Trustworthy AI Agents with Lean 4
10:00-10:45 Web - Muhammad Ahsan Ayaz - Build a Travel agent using Google Gemini, MCP Servers, and Google Agent Development Kit (ADK)
11:45 Break
12:30-13:15 AI - Greg Radzio - Building the Agent Was the Easy Part
13:30-14:15 Mobile - Piotr Prus - Scratch to Reveal: Building a Holographic Card with AGSL, Sensors and Canvas
14:30-15:15 AI - Michał Jabłoński - The Agents Optimization Lab
15:30-16:15 AI - Hrishikesh Yadav - Practical Introduction to DiffusionGemma via Google Cloud
16:30-17:15 Mobile - Renata Wietrzyńska - Mobile A11y First Aid: A Practical Accessibility Workshop for Flutter Developers
17:15-18:00 Break
18:00-18:45 AI - Bianca Ximenes - Scaling AI systems and teams around the world: what actually breaks and why
19:00-19:45 Cloud - Piotr Biniecki - Google Agent Platform in Action: Capabilities, Patterns, and Production Architecture
20:00 Closing
Saturday, 14 November: Talks
7:30 Welcome
8:00-8:30 Mobile - Oleksandr Tilnyi - Trust, But Verify: Certificate Pinning in Flutter
8:35-9:05 Cloud - Ankur Roy - BigQuery, the best non-AI focused AI platform in the world: a deep dive
9:10-9:40 Web - Jakub Andrzejewski - It's Lights Out and Away We Go: Observability, Inspired by Formula 1
9:45-10:15 AI - Paweł Kubiak - From Human-Friendly to Agent-Friendly: WebMCP in Action
10:20-10:50 Robotics - Marcin Wysocki - Three AIs walk into a bar
10:55-11:25 AI - Marcin Czarkowski
11:30-12:00 Web - Tomasz Flis - Beyond Chatbots: Streaming Structured JSON from Gemini to Angular Forms
12:00-12:45 Lunch
12:45-13:15 Web - to be announced
13:20-13:50 Generic - Artur Wojnar - Implementing production-ready systems: DDD is not enough
13:55-14:25 AI - Przemysław Nowak
14:30-15:00 Mobile - Elaine Dias Batista - The State of Voice Interactions on Android
15:05-15:35 Cloud - Pawel Ostromecki - My Barber Shipped While I Debated Architecture: Agentic AI in minutes with Firebase
15:40-16:10 Web - Jarosław Żołnowski - Earn Your Complexity: Designing Angular Systems That Grow With the Product
16:15-16:45 AI - Julia Rapczyńska
16:50-17:20 AI - Tarun Jain - Run On-device AI Applications without internet using Google LiteRT and Qdrant Edge
17:25-17:45 Generic - Mathias Weber - Field Notes of a Technology Scout: What 200+ Tech Events Reveal About Europe's Tech Scene
17:45-18:30 Break
18:30-19:00 Mobile - Dominik Šimoník - Building a Production Video Feed in Flutter
19:05-19:35 Cloud - Çalgan Aygün - Event-Driven AI Pipelines on GCP: From Eventarc to LLMs
19:50-20:15 Closing Ceremony
Details of each keynote, discussion panel, talk and workshop below.
More details will be published soon.
Agenda
Day 1
4:00 PM: Web panel
moderator: Adrian Romański; panelists: Dominika Zajac, Muhammad Ahsan Ayaz, Jarosław Żołnowski, Julia Rapczyńska, Jakub Andrzejewski
16:50: Keynote - Who Let the Dogs Out? The State of AI Robotics - Grzegorz Piotrowski
The robot dogs are out of the lab — and nobody's putting them back. For decades, robotics meant rigidly programmed arms on assembly lines. Now robots finally got the one thing they were always missing: a brain. Vision-Language-Action (VLA) models let a machine hear "clear the table" — and just do it, even on a table it has never seen before.
No robotics PhD required — curiosity is enough. Woof.
17:00: Cloud panel
moderator: Michał Jabłoński; panelists: Goran Minov, Çalgan Aygün, Pawel Ostromecki, Piotr Biniecki, Ankur Roy
17:50: Keynote - RoboPies - Przemysław Nowak
TBA
18:00: Mobile panel
moderator: Artur Skrzypczyk; panelists: Piotr Prus, Elaine Dias Batista, Dominik Šimoník, Oleksandr Tilnyi, Renata Wietrzyńska
18:50: Keynote - Synergy Codes
TBA
19:00: AI panel
moderator: Dawid Perdek; panelists: Greg Radzio, Łukasz Jaźwa, Paweł Kubiak, Marcin Czarkowski, Tarun Jain
19:50: Keynote - Humanoid - Michal Pogoda Rosikon
TBA
Day 2
07:30: Welcome
8:00 AM: Cloud - Goran Minov - Zero Infrastructure, Zero Phishing: Building a Serverless Security Framework on GCP
Commercial spam filters are often "black boxes". Can a developer build a better solution with zero infrastructure?
In this intermediate-level session, I present a custom, event-driven phishing detection framework built entirely on the Google Cloud Platform. We will move beyond high-level concepts and dissect a production-ready architecture where Cloud Run Functions act as the nervous system, Pub/Sub manages the asynchronous data flow, and Gemini provides the intelligence to analyse semantic intent.
This talk is backed by real-world data. I will share the results of this experiment, comparing the "easiness" of the serverless approach against the detection rates achieved. You will see exactly how to chain these services to create a scalable, private security bot that costs pennies to run but delivers impressive results.
Key Takeaways
- How to decouple ingestion and analysis using Pub/Sub and Cloud Run Functions.
- Practical techniques for using Gemini to identify malicious sentiment and phrasing in emails.
- A data-driven look at the cost-effectiveness and accuracy of a custom cloud framework versus traditional methods.
09:00: Robotics - Łukasz Janiec - From Hallucination to Proof: Trustworthy AI Agents with Lean 4
AI agents are increasingly capable of generating and modifying complex software, but in robotics and cyber-physical systems, plausible code is not enough.
In this hands-on workshop, participants will explore how lightweight formal verification can complement conventional development and AI assistance. Using Gemini and other (possibly local models), we will propose invariants, proofs, and repairs for practical robotics problems, while Lean 4 independently checks the corresponding formal claims.
The exercises will cover ROS 2-style lifecycle and resource management, as well as behavior-preserving refactoring of Behavior Trees. Short demonstrations will extend the approach to industrial protocol conformance etc.
The main idea is simple: AI can propose, explain, and repair, but an independent verifier should decide what is accepted as correct.
No previous experience with Lean or formal methods is required.
10:00: Web - Muhammad Ahsan Ayaz - Build a Travel agent using Google Gemini, MCP Servers, and Google Agent Development Kit (ADK)
This codelab is divided in two 🏅 milestones:
A common, simple learning path: setup + instant gratification.
A personalized path where you choose a quest to solve with the assistance of Gemini CLI
The second part makes this codelab perfect for a hackathon (and prizes for the most creative solutions!)
In this particular workshop, you'll learn the following:
- How to get started with Google's powerful Gemini CLI
- How to integrate MCP servers with Gemini CLI
- Creating API keys in AI Studio
- Getting started with Google ADK (Agent Development Kit)
- Providing tools to the AI Agent
- Using Google Search (Grounding) with the AI Agent
- Providing MCP tools (AirBnB MCP server) to the AI Agent as well
By the end of the hands-on workshop, you'll have a working ADK agent that uses the AirBnB MCP server to help as a travel agent, and the powerful Gemini CLI which can be used for a plethora of use cases including coding, searching for information, understand code repositories, and more.
11:45: Break
12:30: AI - Greg Radzio - Building the Agent Was the Easy Part
Building an agent is easy. Making it better requires a repeatable process and evidence that your changes actually help. In this hands-on workshop, we’ll explore the Agent Development Life Cycle (ADLC) through a local assistant agent, using Gemini CLI as our harness of choice. We’ll start in AI Studio to get an API key, connect a Gemini Flash model, and build an assistant around real tasks, MCP tools, and the file system. We’ll shape its context and memory through Skills and a Wiki, and use harness hooks to plug in guardrails that govern how it operates. Then we’ll turn everyday use into a data flywheel: Build → Test → Observe → Refactor. We’ll use Harbor to run task-based evaluations and Google Cloud Trace to inspect execution, understand failures, and identify what needs to change. Each useful failure becomes a test case; each refactor is checked against those tests before the cycle begins again. You’ll learn how to decide whether an agent needs a better tool, clearer instructions, different context, more useful memory, or a stronger guardrail and how to measure whether the change worked. You’ll leave with a practical workflow for continuously improving a local assistant agent, grounded in real tasks, observable behaviour, and repeatable evaluations.
13:30: Mobile - Piotr Prus - Scratch to Reveal: Building a Holographic Card with AGSL, Sensors and Canvas
Realistic materials are among the hardest things to render in a mobile UI. Standard components give you shapes and colours, but a surface that convinces the eye needs light that responds to how the device is held, geometry that responds to the hand, and a way to make those two agree with each other.
This session deconstructs a holographic scratch card built in Jetpack Compose — a small object that ends up needing most of the graphics toolbox at once.
Using AGSL, Canvas and device sensors, we will deep-dive into the compositing rules that decide how independent effects combine into one surface, and the particle work and custom gesture geometry that make an interaction feel physical rather than scripted.
14:30: AI - Michał Jabłoński - The Agents Optimization Lab
Get more from the coding agent you already use. Hands-on workshop. You’ll run short experiments with model choices, better evidence, and delegation to explore the trade-offs in speed, cost, and quality. Bring your usual coding tool and leave with practical techniques you’ve tested-and a clearer sense of when to use them.
15:30: AI - Hrishikesh Yadav - Practical Introduction to DiffusionGemma via Google Cloud
The talk focuses on the Autoregressive model architecture and why there is a need for the decision model in this current scenario of the agentic era and what DiffusionGemma provides, so the aim of mine is to introduce and talk about the experimentation of the decision model as DiffusionGemma and Jev and what one can build with this approach.
Also will have the discussion around how one can implement it well using the Google Cloud.
Agenda for the talk
- Understanding of how LLM works (token generation)
- Why need of the "System one" model, why now ?
- Moving from generating text to making decisions
- LLM as Judge, Useful ? but how reliable ?
- Understanding Jev Approach
- Exploring DiffusionGemma (djev)
- Understanding the Jev inDepth
- Experimentation with the DiffusionGemma and Jev
- Best Approach to make use via Google Cloud
- What future looks like
- Q&A
No Pre-requisite needed for the talk, It takes from beginner to intermediate level
16:30: Mobile - Renata Wietrzyńska - Mobile A11y First Aid: A Practical Accessibility Workshop for Flutter Developers
Accessibility is becoming an increasingly important part of mobile development, especially with the European Accessibility Act bringing more attention to how digital products are built.
However, for developers, the difficult question is often not why accessibility matters - it's where to start and how to make it part of the development workflow.
In this hands-on workshop, you’ll find, fix and verify real accessibility issues in a Flutter app, using the same workflow you can bring back to your team.
We'll start where your users begin: with a screen reader. You’ll learn the essential TalkBack (Android) and VoiceOver (iOS) gestures and experience an app the way a blind or low-vision user does. Then we’ll go through a quick introduction to WCAG 2.1 Level AA, focused only on what really matters in mobile apps.
Next, you’ll get a sample Flutter app from a GitHub repository, intentionally full of accessibility bugs. Together we’ll run a mini audit using manual screen reader testing, then prioritize what needs fixing first.
Then we'll move from audit to implementation. Through practical Flutter examples, we'll look at common accessibility challenges such as buttons, headings and reading order. The most important will be how to verify that a fix actually improves the user experience instead of simply satisfying a checklist.
The goal isn't to memorize WCAG or make accessibility a one-time project. It's to build an A11y First Aid Kit that developers can use throughout the lifecycle of a Flutter application:
Audit -> Prioritize -> Fix -> Test -> Document -> Repeat
What to bring:
- A laptop with Flutter SDK installed
- An Android or iOS phone with developer mode enabled (recommended for screen reader testing)
- Headphones, so you can hear your screen reader without disturbing others
Level: Beginner to intermediate. Basic Flutter knowledge required, no prior accessibility experience needed.
17:30: Break
6:00 PM: AI - Bianca Ximenes - Scaling AI systems and teams around the world: what actually breaks and why
Scaling a distributed data and AI org, growing from a handful of people to tens across time zones, running systems that serve millions of requests, breaks in the exact place agentic AI systems break: the gap between a guideline and an enforced rule. Distributed teams and agentic systems fail the same way for the same reason. This talk maps what actually breaks when that gap opens, architecturally (systems no one owns end to end), organizationally (central versus federated authority), and in governance (rules that exist on paper but nothing enforces), with the trade offs of each fix and what to prioritize first.
https://arxiv.org/abs/2606.20762
19:00: Cloud - Piotr Biniecki - Google Agent Platform in Action: Capabilities, Patterns, and Production Architecture
Google’s Agent Platform - built on the Agent Development Kit (ADK) and Google Agent Runtime - provides the managed infrastructure required to elevate AI agents from standalone prototypes into production-grade multi-agent systems. But what do these platform primitives actually look like when deployed against complex enterprise data?
In this 45-minute technical walkthrough, attendees will explore the core capabilities of the Google Agent Platform through a live, production-grade reference architecture: the Marketing Agentic Team Extension (ATE), built internally by Sii experts. ATE unifies fragmented data across Google Analytics 4, Google Ads, Google Tag Manager, and BigQuery using a coordinated mesh of specialist agents.
20:00: Closing
Official Closing Ceremony
Day 3
07:30: Welcome
8:00 AM: Mobile - Oleksandr Tilnyi - Trust, But Verify: Certificate Pinning in Flutter
Making Flutter a solid choice for fintech is a challenge. Choice which is not exclusively about reliability, performance and maintenance cost. In high-compliance environments - a penetration test is reckoning.
This talk goes beyond ‘use HTTPS’. During this you'll discover the certificate validation hooks Dart exposes and how certificate pinning lets you define exactly who your app trusts, making that trust explicit and verifiable. We will walk through implementation strategies for certificate pinning in Flutter - from the dead-simple to the enterprise-grade.
You'll leave with a decision framework you can defend in your next security review - not just working code, but the reasoning behind it.
Key takeaways:
- Certificate pinning strategies available in Flutter
- SPKI pinning vs full-certificate pinning
- Decoupling security from your app release cycle
08:35: Cloud - Ankur Roy - BigQuery, the best non-AI focused AI platform in the world: a deep dive
BigQuery in GCP would be successful even if it didn't have AI, but it wouldn't be as accessible, nor as flexible. Join a deep dive into the BigQuery service itself (which is the service of choice for companies like Spotify and HSBC long before AI) and a dive into all the AI services that have been added in order to faciltiate BigQuery over the years. This deep dive will go through the concept of BigQuery and the AI that now supports it (even agents!).
09:10: Web - Jakub Andrzejewski - It's Lights Out and Away We Go: Observability, Inspired by Formula 1
Formula 1 teams don’t rely solely on the driver’s feedback—they rely on telemetry. Every lap, every corner, and every millisecond is measured to understand what really happened on the track. Modern Web applications deserve the same approach. In this talk, we’ll explore how to build an observability strategy inspired by Formula 1, moving beyond synthetic benchmarks to real user experience.
09:45: AI - Paweł Kubiak - From Human-Friendly to Agent-Friendly: WebMCP in Action
For the last few decades, we’ve been building the web primarily for humans. We’ve improved user experience, introduced better design practices, and made accessibility a fundamental part of modern web development. But now, a new type of user is starting to interact with our applications: AI agents. Today, agents can navigate websites by analyzing screenshots, reading the DOM, or using the Accessibility Tree. These approaches work, but they can also be expensive, inefficient, and sometimes unpredictable. What if, instead of forcing agents to understand our user interfaces, web applications could explicitly tell them what they are capable of? This is where WebMCP comes in. In this session, we’ll explore how WebMCP can expose application capabilities to AI agents as structured tools, allowing them to interact with the web in a more intentional and reliable way. We’ll also look at the bigger picture: where these agents can actually live - embedded directly in a web application, inside the browser, or even completely outside of it. Through practical experiments and live demos built with Angular, WebMCP, browser AI capabilities, and external agents, we’ll see what this new model of interaction could mean for web developers. The web has spent decades becoming more human-friendly. Now it might be time to make it agent-friendly as well.
10:20: Robotics - Marcin Wysocki - Three AIs walk into a bar
One knows what a bar is. One can dream about ordering something. One can actually walk. None of them can do all three, and that has been the state of robotics... but that's beginning to change.
We watched language models explode and transform how we work. Physical intelligence is approaching its own GPT moment, with humanoids as the headline act and quadrupeds as the opening band that got there first. Every serious robotics lab is now introducing these three AIs to each other, and it's all beginning to move faster and faster.
This talk walks through the stack behind that moment. You'll meet the vision-language model that knows what everything is but has no body, the world model where a robot can dream through a thousand attempts before risking one for real and the reinforcement-learned policy that fell over two billion times in simulation before it could walk. You'll learn how the field is wiring them together with VLAs and what comes after. What already runs on real hardware, what is still demo-ware and why the near future of robotics feels less like programming and more like raising something.
10:55: AI - Marcin Czarkowski
11:30: Web - Tomasz Flis - Beyond Chatbots: Streaming Structured JSON from Gemini to Angular Forms
While AI chatbots are ubiquitous, enterprise applications demand structured, reliable data entry rather than raw conversational text. In this talk, we will explore how to build an intelligent, real-time form copilot that extracts messy user inputs into validated form fields on the fly.
We will cover the full pipeline: enforcing strict JSON Schemas with the Google GenAI SDK (Gemini / Vertex AI), consuming token streams in real time using partial JSON parsing, and patching Angular Reactive Forms efficiently. You'll learn how Angular Signals, fine-grained reactivity, and Zoneless execution allow you to handle high-frequency AI streams without triggering performance bottlenecks or UI lag.
12:00 PM: Lunch
12:45: Web - Sam Vloeberghs - Angular × AI - the framework goes AI-native
AI is rapidly changing the way we build software, and Angular development is no exception. With an ever-growing ecosystem of AI-powered tools, framework integrations, and community projects, it's becoming easier than ever to build, understand, and maintain Angular applications with AI as your development partner. In this session, we'll explore the current landscape of AI tooling for Angular, from official initiatives by the Angular team to the most useful community-driven solutions. You'll see what's available today, how these tools fit into real-world development workflows, and leave with a practical overview of how to build Angular applications more effectively with AI.
13:20: Generic - Artur Wojnar - Implementing production-ready systems: DDD is not enough
DDD helps us manage complexity in the business domain, but some of the hardest problems emerge between modules and services - where communication becomes asynchronous and unreliable.
Messages get duplicated, events end up forgotten in dead-letter queues, consumers fail halfway through processing, and traffic spikes expose hidden assumptions. And sometimes everything works perfectly from a technical perspective, yet nobody uses the feature we built.
This talk explores what it takes to move from a well-designed system to one that can be safely operated in production.
What will you learn?
- How to design reliable asynchronous communication using queues and logs.
- How to handle lost and duplicated events with patterns such as Inbox and WAL.
- How to use monitoring, observability, alarms, and quality metrics to detect problems before users do.
- How to define success metrics and verify whether your features actually deliver value.
13:55: AI - Przemysław Nowak
14:30: Mobile - Elaine Dias Batista - The State of Voice Interactions on Android
Android has provided Speech Recognition APIs and Voice Actions for several years. Then, the Google Assistant came in and allowed us to integrate with it using App Actions. Also, several third party solutions exist and can work not only for Android but also for iOS, the web and IoT devices. In the era of Cloud APIs, Gemini, LLMs, GenAI and privacy concerns, are voice interactions easier, more interesting and more useful for our users? Or is it the other way around? What new challenges are we facing and how can we solve them? Come to this talk and we'll go together through the exciting path of voice interactions on Android!
15:05: Cloud - Pawel Ostromecki - My Barber Shipped While I Debated Architecture: Agentic AI in minutes with Firebase
Abstrakt: Every time I start a side project, I catch myself doing the same thing. I pick a technology I want to learn, build far more structure than the project needs, and somewhere along the way lose sight of the problem I set out to solve. A friend of mine runs barbershops and a haircare brand and has never written code. He knows his business inside out, described his problem to an AI, and had something running while I was still choosing patterns.
In this session I will demo an app solving a real problem from his business, built the second way, taking full advantage of what Firebase already gives you instead of rebuilding it:
- Angular on the frontend
- Firebase SQL Connect, with Cloud SQL for PostgreSQL underneath
- pgvector for retrieval
- Gemini for reasoning
- Firebase handling auth, schema, hosting, and deployment
Then we look at what changes once the app has to run at scale, and where the questions stop being about building and start being about operating.
Domain knowledge is what gets a product built today. Keeping it alive in production is where the engineer is still irreplaceable.
15:40: Web - Jarosław Żołnowski - Earn Your Complexity: Designing Angular Systems That Grow With the Product
What if your architecture didn’t have to be complicated on day one?
We often start architecture discussions with the tools: Should we use Nx? SignalStore? SSR? WebSockets? Microfrontends? But what if we started somewhere else - with the product, the people using it, and the problems we can actually measure?
In this talk, we’ll explore the architecture of a consumer photo-sharing app with Angular, starting from the product rather than the technology. Through a practical case study, we’ll walk through the decisions behind the architecture and, more importantly, the signals that tell us when a simple solution is no longer enough.
You’ll leave with a practical framework for making architectural decisions in your own Angular applications - knowing not only what to introduce, but when you’ve actually earned it.
16:15: AI - Julia Rapczyńska
16:50: AI - Tarun Jain - Run On-device AI Applications without internet using Google LiteRT and Qdrant Edge
Many AI applications still depend on cloud APIs for retrieval and inference, making them difficult to use in environments with unreliable connectivity or strict privacy requirements. This session explores how to build AI applications that run entirely on device using Qdrant Edge for in process vector search and Google's LiteRT to run models such as Gemma 4 locally. You will learn how to combine local retrieval and inference into a single pipeline that works without an internet connection while keeping data on the device.
17:25: Generic - Mathias Weber - Field Notes of a Technology Scout: What 200+ Tech Events Reveal About Europe's Tech Scene
From discussions about custom system-on-chip design at an Amazon tech event in Gdańsk, to conversations about agentic systems in a technology consultancy near the Port of Helsinki; from a vibe-coding session hosted by a global pharmaceutical company in Berlin, to debates in Kraków about the convergence of psychedelic neuroscience and artificial intelligence; and from there to an exchange with the co-founder of an AI lakehouse platform, the European technology scene is anything but quiet. Behind the polished keynotes and carefully rehearsed product demos, engineers, founders, researchers and technology leaders are wrestling with the same questions: Which ideas are becoming engineering reality? Where does the hype end? What is actually working in production? How do organizations adapt? In this session, technology scout and data architect Mathias Weber opens his field notes from more than 200 technology events across Europe, with a particular focus on Berlin’s vibrant technology community. Drawing on these observations, he shows how technology scouting can help identify relevant signals, evaluate emerging trends and support better decisions in a rapidly changing environment.
17:45: Break
18:30: Mobile - Dominik Šimoník - Building a Production Video Feed in Flutter
Video playback in Flutter looks simple until you ship it. The framework gives you the building blocks, but a production-quality video feed demands answers to questions you didn't know to ask: How many native players can a device actually run? What happens when you pause and resume a buffered stream? Why does your app crash on the 20th swipe with no error? We'll share the patterns we built for Divine, a short-form looping video app, from player pooling and muted preloading to synchronous caching and platform-specific workarounds. Each one came from a real production problem, and together they form an architecture you can adapt to your own app.
19:05: Cloud - Çalgan Aygün - Event-Driven AI Pipelines on GCP: From Eventarc to LLMs
In this live building session, I will create a production-ready, event-driven AI pipeline on Google Cloud from a blank terminal to a working system.
A file lands in Cloud Storage, Eventarc routes the event, Cloud Run processes it reliably, and Vertex AI with Gemini enriches the data through classification, summarization, and entity extraction. The enriched results are then stored.
The focus is not on AI hype or a pre-recorded happy-path demo. Attendees will see the system being built live, including the practical decisions, trade-offs, and failure scenarios involved in making it production-ready.
19:50: Closing Ceremony
---
Speakers
Elaine Dias Batista - SFEIR (Senior Staff Engineer & Google Developer Expert)
Elaine is a Senior Staff Engineer at SFEIR where she works on projects including mobile technologies since 2013. She has worked on several projects in native Android/iOS and Flutter. She's a GDE for the Google Assistant, Flutter and Dart.
Paweł Kubiak
Dominik Šimoník
Dominik is a software developer, passionate about Cloud technologies and mobile app development. He is a GDG community events organizer at GDG Ostrava (GUG.cz).
Oleksandr Tilnyi - Devinti (Tech Lead | Senior Flutter Engineer)
With over 6 years of experience, Oleksandr brings a strong engineering foundation shaped by working across a variety of projects - from dynamic startups to large mobile systems. He has consistently contributed as a core team member, helping enforce best industry practices and delivering scalable, maintainable code with a strong focus on readability and long-term value.
His background in…
Ankur Roy - Online Partner (Google Developer Expert • Solutions Architect • Author)
Solutions Architect at Online Partner AB. Ankur worked on several projects involving Google Cloud services and Google Workspace along with multiple other technology stacks and platforms.
Jakub Andrzejewski
Marcin Wysocki
Marcin Czarkowski
Tomasz Flis
Sam Vloeberghs - Innovatus (Software Engineer)
Sam is a freelance software engineer and Internet entrepreneur, currently focusing on frontend technologies. Co-organiser of NG-BE and Angular Belgium meetup.
Personal Website: https://samvloeberghs.be
LinkedIn: https://www.linkedin.com/in/samvloeberghs/
Przemysław Nowak
Pawel Ostromecki
Jarosław Żołnowski
Julia Rapczynska
Tarun R. Jain - Google Developer Expert, AI (Founding Engineer)
Founding Engineer, GDE in AI and YouTube Content Creator at AI with Tarun
Mathias Weber
Hosted By
Luka Malakhau, Software Developer
Karol Wrótniak, GDG Organizer
Experienced Android & Flutter Developer, Public Speaker, and Technical Writer sharing insights on mobile development, Open Source maintenance, and modern CI/CD workflows.
Artur Skrzypczyk, Organizer
Jan Łuczka, Android Developer
Szymon Mazanik, Flutter Lead
Adrian Romański, Organizer
I’m a Full Stack Developer working primarily with Java and Angular, with a strong focus on modern frontend development.
Over the past years, Angular has become my main area of specialization, both in day-to-day work and continuous self-development.
I’m particularly interested in:
- Angular architecture, performance, and component design
- State management using RxAngular and NgRx
- Nx-based monorepo architectures for scalable applications
- Web Accessibility (a11y)
- Developer Experience and tooling
- AI-powered solutions supporting frontend and developer productivity
Lately, I’ve been diving deeper into accessibility, focusing on semantic HTML, WCAG guidelines, and building inclusive, accessible Angular components.
Dawid Perdek, Organizer
Software Developer with a decade of experience in various technologies, company sizes and industries. Focus mostly on frontend. Curious of how products are making money, loves good food and good people.
---
Partners
Kolo Naukowe Solvro (https://solvro.pwr.edu.pl/pl)
Machinekind (https://machinekind.ai)
Uczymy maszyny poruszać się w świecie ludzi.
Kolektyw inżynierów uczenia maszynowego, robotyków i projektantów. Budujemy warstwę inteligencji razem ze sprzętem, na którym da się ją uruchomić.
JS Poland (https://js-poland.pl)
The Biggest JavaScript Conference in CEE
NG Poland (https://ng-poland.pl)
The Angular Community, Together on the Big Screen
Synergy Codes (https://www.synergycodes.com)
Synergy Codes is the specialist studio for performant, production-grade diagram interfaces.
Since 2011 we have been building on technology and IP we own and maintain ourselves – four products and a 15-year delivery track record – and shipping visual systems for clients across the full spectrum: innovative startups validating a first product, mature SaaS platforms scaling against thousands of concurrent users, and Fortune 500 enterprises with mission-critical workflows.
inQUBE (https://inqube.pl)
Wrocław University of Economics (https://uew.pl)
AI HUB @ UEW (https://ai.uew.pl)
Łączymy naukę, biznes i edukację wokół sztucznej inteligencji
---
Complete your event RSVP here: https://gdg.community.dev/events/details/google-gdg-wroclaw-presents-google-devfest-2026-wroclaw/.
