DevFest Budapest 2025 - The AI Conference

DevFest is the world's largest community-driven tech conference, bringing together passionate developers from around the globe. The DevFest Budapest Developer Conference offers a unique opportunity to explore developer tools, learn from young researchers, local, and international industry experts, and connect with fellow developers.

🚀 Go All-In on AI: From Agents to Production

This year, DevFest Budapest is moving beyond the hype and diving deep into the practical future of software: Artificial Intelligence. Our 2025 conference is hyper-focused on the technologies that are actively reshaping development.

We are bringing together Google Developer Experts, industry leaders, and academic pioneers to share hands-on, practical knowledge. Get ready for a day packed with deep dives on:

Building AI Agents: Go from zero to a functioning agent. We'll explore multiple agentic frameworks, including Google's Agent Development Kit (ADK), and even build one live on stage.
GenAI on Google Cloud: Master the full stack, from deploying open-source Gemma models on Cloud Run to architecting complex, serverless multi-agent systems with Vertex AI.
AI in Practice: See how Gemini is revolutionizing DevOps, how LLMs are solving real-world business challenges, and how AI is powering new frontiers in bioinformatics.

Whether you're a student eager to deploy your first GenAI app or a professional developer architecting the next wave of intelligent systems, you'll leave with the practical skills and expert connections to build what's next.

The DevFest Budapest Developer Conference is organized as a collaboration of Department of Automation and Applied Informatics of Budapest University of Technology and Economics (BME), GDG Budapest, GDG on Campus ELTE, GDG on Campus BME, and Zenitech.

The DevFest Budapest Developer Conference will be hosted at the Budapest University of Technology and Economics (BME), in BME Building "I".

Sign up now to reserve your seat!

---

A DevFest a világ legnagyobb, közösség által szervezett technológiai konferenciája, amely összehozza a fejlesztőket, a világ minden tájáról. 

A DevFest Budapest Fejlesztői Konferencia kiválló lehetőség a legújabb fejlesztő eszközök megismerésére és a többi fejlesztővel való kapcsolatteremtésre. Fiatal kutatók és ipari szekértők kiválló előadásaival várunk!

A DevFest Budapest Fejlesztői Konferencia a BME Automatizálási és Alkalmazott Informatikai Tanszék, a GDG Budapest, GDG on Campus ELTE, GDG on Campus BME és a Zenitech közös szervezésében kerül megrendezésre, a Budapesti Műszaki és Gazdaságtudományi Egyetem (BME) "I" épületében.

Regisztrálj, és biztosítsd a helyed!

Agenda

Day 1

2:00 PM: ARRIVAL, REGISTRATION

2:30 PM: Conference opening

Welcome speech

2:40 PM: Navigating Agentic Frameworks: Practical guidance for developers - Máté Sal (BlackRock)

Agentic AI frameworks are transforming how we build intelligent systems by enabling LLMs to plan, act, and collaborate autonomously. In this talk, we’ll compare five leading frameworks—LangGraph, CrewAI, OpenAI Agents SDK, Google ADK, and AutoGen—exploring their mental models, strengths, trade‑offs, and real‑world use cases. You’ll leave with a clear decision framework and practical insights to help you choose the right tool for your next project.

3:10 PM: The 30-Minute Agent: A Hands-On Live Build - Gábor Volczér (Aliz)

Everyone's talking about AI agents, but how fast can you actually build one? In this fast-paced, 30-minute live-coding session, we're skipping the slides and building a functional AI agent from scratch. You'll see the Gemini CLI, combined with Google's Agent Starter Pack for rapid bootstrapping, and the Agent Development Kit (ADK) in action. By the end, our agent will be live, deployed on an "Agent Engine." Get ready, it's going to be fast.

4:00 PM: Agents to Zero: Experiments with Serverless Multi-Agent Architectures on Google Cloud - Akshata Mohanty (Google Developer Expert)

Anyone can build an agent, but building the right agent is still a challenge. Prototyping speed, deployment, and unit economics are critical. Using a real-world demo, this talk progressively walks through the lifecycle of an AI product: from leveraging Vertex AI to deploy a simple inference-as-a-service model server to building a complex multi-agent serverless architecture using ephemeral, containerized, stateless agents on Cloud Run. The talk will briefly cover concepts such as host-remote agent patterns with ADK, agent communication protocols, stateless memory management and handling known serverless bottlenecks such as cold starts, latency and concurrency. Attendees will see how agents instantiate, collaborate, and scale to zero, uncovering practical insights into building elastic, cost-efficient, event-driven AI systems in the cloud.

4:30 PM: Latest advancements in intelligent data processing and AI technologies - Gergely Major (Zenitech)

This talk explores the latest trends and advancements in intelligent data processing and AI technologies, with a special nod to how these innovations are reshaping modern production management. We’ll take a broad look at how data, automation, and smart solutions are changing the way we make decisions and optimize complex systems. From predictive analytics to AI-driven process control and LLMs, the goal is to spark ideas and discussion rather than dive too deep into any single topic. It’s an open, forward-looking session for anyone curious about where intelligent technologies are heading next.
The talk presents results created in the projects named "AI-assisted production management / Mesterséges intelligenciával támogatott termelésirányítás, code: 2019-1.1.1-PIACI-KFI-2019-00263

5:20 PM: Predicting Customer Intents - Szabolcs Domján (Deutsche Telekom)

Many of us have experienced how difficult it can be to navigate through the maze of options to reach the right support group. Customer Intent Prediction enables us to know the reason for a customer’s contact before they say a word. By using LLM-based models, we turn static dialog into a personalised conversation, offering customised recommendations from the start to reduce time to resolution and improve first call resolution rates. Join us to see how the Google Cloud stack is changing the way we interact with our customers.

5:50 PM: How AI is Reshaping Software Delivery Inside Remote Engineering Teams - Adam Drobinoha (Adroit Group)

This talk explores how AI is transforming the way remote software development teams work and what clients now truly value from engineering partners. Instead of simply delivering hours of coding, teams are shifting toward solving problems, shaping products, and making smarter architectural decisions. We’ll look at what has changed inside day-to-day development work, how expectations are evolving, and what organizations can do to stay competitive. The goal is to show a practical and realistic path forward, not hype.

6:50 PM: From Pixels to Crops: Building Practical AI for Agriculture - Elene Samsiani (ELTE)

AI in agriculture isn’t just a concept, it’s something we’ve tested, built, and improved in real conditions. In this session, I’ll share hands-on lessons from developing an AI system that uses computer vision and drone-based data collection to identify agricultural pests and support crop management. Our team gathered practical experience by visiting fields, capturing and labeling images, and learning how unpredictable natural environments can challenge even the most advanced models.

I’ll discuss how we trained and optimized CNNs for agricultural image classification, integrated drone imagery, and explored ways to combine these insights with large language models (LLMs) for clearer, more accessible pest management guidance. The project was recognized at a national innovation competition in Georgia, reflecting the value of building AI that works not just in theory but in the field.

Attendees will gain a realistic view of how to bring AI from lab to land, blending computer vision, LLMs, and hands-on testing to create responsible, sustainable, and practical technology for agriculture.

7:20 PM: Comics with Nano Banana - Bálint Décsi, Abdulazizbek Gainazarov (AI Storybook)

A practical architecture for building web apps on Firebase and staying in Google Cloud ecosystem using serverless solutions with containers, Functions, Storage and Firestore. This is our way of multimodal comics generation using Nano Banana, avoiding text quality deterioration and keeping consistency throughout panels.

Day 2

5:20 PM: Cloud and Open-Source GenAI for Beginners: Deploy Your First LLM App on Google Cloud Run - Bobur Yusupov (ELTE)

Generative AI is transforming how we build applications, but getting started can feel overwhelming for beginners. In this session, I’ll guide you through deploying your first open-source LLM app using Google Cloud Run, making cloud-native GenAI accessible and hands-on.

Attendees will learn how to deploy open-source models like Gemma 3 to Cloud Run, and expose it as a simple GenAI API. The talk emphasizes practical, step-by-step instructions, making it suitable for developers, students, and hobbyists with little to no prior experience in cloud or AI deployments.

I’ll also cover essential concepts such as model selection, lightweight inference, and scaling options, giving beginners a clear roadmap from experimentation to a functioning cloud app.

By the end of the session, participants will have the confidence and knowledge to deploy their own LLM-powered GenAI applications, understand the basics of cloud-native AI architecture, and explore further experimentation with open-source models. This session is designed to make AI approachable, practical, and beginner-friendly.

5:50 PM: AI Meets Bioinformatics: Transforming Life Sciences from Data to Discovery - Latifa Lechlech (Sorbonne University)

Biology is entering its computational era. As datasets grow from gigabytes to petabytes, artificial intelligence is redefining how we discover drugs, design proteins, and personalize healthcare. The convergence of bioinformatics and AI isn’t just accelerating research, it’s reshaping the entire life sciences ecosystem.

This talk explores how the convergence of bioinformatics and AI is transforming the way we understand life, disease, and therapeutic innovation. We’ll begin by revisiting the foundations of bioinformatics from the analysis of genomic, proteomic, and clinical datasets and the growing challenge of managing biological "big data". Then, we’ll see how AI and deep learning models such as AlphaFold 3 and ESM2 have revolutionized structural biology by predicting protein-DNA, RNA, and ligand interactions with atomic precision.

Moving from molecules to medicine, we’ll examine how AI-driven drug discovery, multi-omics integration, and precision healthcare are enabling personalized treatment strategies and accelerating biomedical research.

Finally, we’ll discuss the next frontier: generative AI for biology, ethical considerations, and the need for interdisciplinary collaboration between life scientists and computational experts. Whether you’re a biologist, data scientist, or healthcare innovator, this session offers a forward-looking perspective on how intelligent algorithms are reshaping the future of biological discovery.

6:50 PM: DripCheck AI - Kendi J (ELTE)

DripCheck AI, is a hands-on Flutter codelab where you’ll build a smart fashion assistant that tells you if your outfit “drips.” Powered by Google’s Gemini, this app turns simple text prompts like “white tee, denim jeans” into curated style recommendations complete with complementary colors and accessory suggestions. No prior AI experience needed; by the end of this session, you’ll have a fully working Flutter app integrated with an AI backend, plus the know-how to adapt it to any prompt-driven use case.

7:20 PM: Using Gemini for devops tasks - practical example with helm-chart - Alexander Grigoryev (Kubernetes Specialist)

In this talk, Alexander will discuss how Gemini 2.5 Pro in Copilot can be useful when programming and modifying Helm charts (package manager) in Kubernetes (k8s). He'll also discuss what you need to know when working with it and how to effectively validate the results.
We'll explore a real technical issue at a client, its first temporary workaround, and subsequent solutions using GitHub Copilot in agent and advisor (ask) mode. We'll analyze the pros and cons of both approaches and arrive at the right technical solution. During the process of research and implementation, not a single cluster crashed.

---

Speakers

Akshata Mohanty - Google Developer Expert in AI/ML (Head of Engineering - Tilla)

Akshata is a software engineer and cloud architect. She is a Google Developer Expert (GDE) in AI/ML.
Currently, Akshata leads engineering at Tilla Technologies in Berlin. In 2024, she founded Keis in Singapore, focusing on helping companies launch generative AI products on Google Cloud.
Akshata graduated from the Indian Institute of Technology (IIT) Kharagpur in India. She is an active t…

Máté Sal - BlackRock (Lead AI Engineer)

Mate Sal is a Lead AI Engineer on the Aladdin AI Platform Team at BlackRock, driving the development of Responsible AI features to enable scalable and safe AI solutions across the company.

Gábor Volczér - Aliz (ML and Data Engineer)

For the past decade, I have been deeply engaged in delivering Google Cloud solutions across a diverse range of projects. My expertise lies primarily in the realm of Big Data, where I have successfully navigated complex challenges and implemented innovative solutions. Recently, I have expanded my focus to encompass the burgeoning field of Artificial Intelligence, leveraging my technical prowess…

Gergely Major - Zenitech (AI Developer specializing in Machine Learning)

I'm an AI developer at Zenitech Consulting Zrt., specializing in computer vision and LLM-driven agents. I’m passionate about bringing ideas to production and creating impactful, real-world solutions.

Adam Drobinoha - Adroit Group (CEO)

Adam is the CEO of Adroit Group, a fully remote software development company based in Budapest. He has led teams delivering complex platforms for global clients in fintech, trading, and operations. His work focuses on combining strong engineering practices with practical business problem-solving. He’s currently exploring how AI changes the way distributed teams collaborate, deliver value, and …

Szabolcs Domján - Deutsche Telekom IT Solutions Hungary (AI Product Manager - Data Tribe, Head of Team Fixed Access Analytics)

My mission is to democratize decision intelligence and safe, reliable AI. I bring a unique combination of deep technical expertise, product mindset, and the ability to build simple and effective solutions to complicated problems from conception to completion. Passionate about leading product teams by focusing on outcomes and data insights to solve user problems.

Bálint Décsi - Deutsche Telekom IT Solutions Hungary (Data Engineer)

Graduated with engineer titled degree. Actively doing side projects. Data-domain practical knowledge in telco companies and developing geospatial analytics solutions on Google Cloud. Having had experience in the research field in deep learning for biomedical imaging. Having good communication skills in various languages. Teamwork and project lead experience during my time at mesh., creating a …

Alexander Grigoryev

Alexander is a Devops Engineer with big passion and interest in Kubernetes.
He is also, an entrepreneur and a good cyclist.

Latifa Lechlech - Sorbonne University (Master's Graduate in Bioinformatics and Modeling)

Latifa has Master’s graduate in Bioinformatics and Modeling from Sorbonne University and Artificial Intelligence from Eötvös Loránd University. Her research interests focus on the intersection of computational biology and artificial intelligence, particularly the application of machine learning and graph neural networks (GNNs) to molecular representation, drug discovery, and protein-protein in…

Elene Samsiani - Eötvös Loránd University (Co-founder of AI-driven innovation projects)

Elene Samsiani is a young innovator and cybersecurity intern passionate about AI and emerging technologies. She has experience in building and managing projects that combine artificial intelligence and real-world applications, including developing a computer vision model for mobile use and leading awareness initiatives in tech education.
Elene has been a finalist in several national and …

Kendi J. - Eötvös Loránd University (Software Engineer)

I'm a software engineer who enjoys solving problems and building cool things.
When I'm not working, I love soaking in the beauty of mountains and staying connected with awesome tech communities =)

Bobur Yusupov - Eötvös Loránd University (Student, DevOps Engineer Intern at Ericsson)

Bobur, is a Computer Science student at Eötvös Loránd University (ELTE) and a DevOps Engineer Intern at Ericsson, where he focuses on cloud infrastructure with AWS and Azure. Previously, he worked as a Software Developer Intern at Silk Road Professionals. He was a participant in the Azimi Fellowship, a program organized by zypl.ai to provide practical knowledge in Machine Learning and statisti…

Abdulazizbek Gainazarov - Eötvös Loránd University (CS Student)

I am your fellow from ELTE: 3rd year Bachelor in Computer Science, work in WizzAirr for almost 2 years, and at meantime build a startup in mesh.

Hosts

Péter Ekler Dr. - Budapest University of Technology and Economics (Associate Professor)

Péter Ekler Dr. is Associate Professor at the Budapest University of Technology and Economics - Faculty of Electrical Engineering and Informatics, Department of Automation and Applied Informatics.

Bertalan Forstner Dr. - Budapest University of Technology and Economics (Deputy Department Head and Associate Professor)

Bertalan Forstner Dr. is the Deputy Department Head and Associate Professor at the Budapest University of Technology and Economics - Faculty of Electrical Engineering and Informatics, Department of Automation and Applied Informatics.

Zsigmond P. Toth

Zsigmond is the organizer of the GDG Budapest and GDG Cloud Budapest communities and the DevFest Budapest and DevFest Live events. He is a Google Cloud Certified Professional Cloud Architect and he works as a Systems Architect at EPAM Systems in Budapest, Hungary.

Akbermet Toktobekova

Akbermet Toktobekova is a Computer Science student at ELTE University and Google Developers Group on Campus ELTE Organizer. She has experience in AI integration for business workflows and in organizing various tech and community events.

Adil Zhetigenov

Adil Zhetigenov is a Computer Science student at ELTE University and Google Developers Group on Campus ELTE Organizer. He has experience in business workflow automation and event organisation for tech communities

Aamna Tayyab - Adroit Group (Head of Delivery)

Aamna is the organizer of the GDG Budapest and the DevFest Budapest event. She is a founder & pioneer of GDSC ELTE on campus community.
She works as the Head of Delivery at Adroit Group, responsible for leading cross-functional teams and ensuring successful execution of technology initiatives across the organisation.

---

Partners

GDG on Campus Eotvos Lorand University (https://gdg.community.dev/gdg-on-campus-eotvos-lorand-university-budapest-hungary/)
GDG on Campus Eotvos Lorand University - Budapest, Hungary is an independent group; our activities and the opinions expressed here should in no way be linked to Google, the corporation.

Women Techmakers Budapest (http://womentechmakers.hu)
Women Techmakers Budapest is the first WTM group in Hungary. Women Techmakers provides visibility, community, and resources for women in technology to drive participation and innovation in the field.

---

Complete your event RSVP here: https://gdg.community.dev/events/details/google-gdg-budapest-presents-devfest-budapest-2025-the-ai-conference/.