Agenda

Domain-Driven Transformation—Modularizing and Modernizing Legacy Software

Future of Engineering Stage — 45 minutes Future of Engineering Clean Architecture Legacy Software

Starting a new project from scratch with a clean slate is one thing; tackling legacy systems with chaotic architectures is another. Many teams inherit systems described as “Monoliths” or “Big Balls of Mud,” and this talk focuses on how to bring order to such challenges.
Participants will learn how to evaluate their current systems and identify steps to introduce—or restore—structure. Using Collaborative Modeling, attendees will gain actionable insights to design a clear and future-ready architecture. The talk will equip participants with practical strategies to address common challenges, such as BBOM architecture, anemic domain models, and disorganized teams. By the end of the session, attendees will have a catalog of proven refactorings, actionable heuristics, and concrete tools to immediately evolve their own projects. Read More...

• Henning Schwentner [Coding Storyteller — WPS]

From Copilot to Cowork: Software Engineering in the Age of AI Agents

Future of Engineering Stage — 30 minutes Software engineering AI Agents

Software engineering is moving beyond AI-assisted development toward agentic systems—AI that can reason, plan, use tools, and execute multi‑step work.  Join Bogdan Stelea, Senior Software Development at Microsoft, as he presents a practical, engineering-first view of GenAI-driven, agentic software development. The talk will cover the ways in which AI agents move software from deterministic execution to intent-driven, non‑deterministic systems, while still meeting production requirements for reliability, security, and scale.
Key Topics CoveredAgentic-oriented software engineeringDesigning and integrating AI agents into production systemsHow agentic AI reshapes the SDLCVerification, validation, and observability for non-deterministic behaviorSecurity and Responsible AI by design
Who Should AttendSoftware engineers and architects building or exploring GenAI-powered systemsTech leads and engineering managersTech enthusiasts interested in finding out more about the future operating model of software engineering Read More...

• Bogdan Stelea [Senior Software Engineer — Microsoft]

Goodbye Legacy, Hello Lakehouse: The Databricks Migration Story

Future of Engineering Stage — 30 minutes

• George Bărbulescu [Head of Data for Digital — BCR]

Is the Specialist Dead? Being a software developer in the AI age

Future of Engineering Stage — 45 minutes Artificial Intelligence Future of Engineering Clean Architecture Software Development

My journey in tech began with a Commodore 64. From there, I followed the industry's gravity: moving to the PC, finding a home on the Mac, and building a professional career through the mobile revolution—from the early days of feature phones and BlackBerry to the heights of Android.  We moved from direct memory access on those ealy machines to various architecture and abstractions.More recently, I transitioned into Full Stack development, seeking a broader view of the systems we build. It felt like the right moment. The world of engineering is changing drastically. Do we even still need the traditional split of specializations like mobile, frontend and backend developers? Grady Booch recently noted, we have entered the 3rd Golden Age of Software Development. I remember studying his software modeling techniques back at university.Is this the Renaissance of the Software Architect as we are moving past the "move fast and break things" era? Read More...

• Danny Preussler [Engineer Integrations & Tech Lead Internal Tools — SoundCloud]

Lunch Break

Future of Engineering Stage — 1 hour

APIs for the Age of Agents

Future of Engineering Stage — 45 minutes Future of Engineering APIs

APIs have always been about exposing capabilities — but who's consuming them is changing fast. Alongside human developers, we now have two new types of API consumers: runtime agents that call APIs autonomously to get things done, and agentic coding tools that help developers and engineers build integrations faster than ever before. This shift changes what "good API design" means. Beyond readability for human developers, APIs now need to be optimized for machine consumption — what we might call agent experience. And above the API layer, workflows offer a more powerful abstraction: goal-oriented interfaces that let agents (and the people building with them) work at a higher level than individual endpoint calls.Making this work safely at enterprise scale requires two things: a security layer that brokers credentials so agents never hold secrets directly, and sandboxed environments that let agents and builders operate freely without risk to production systems.If you've already invested in your API estate, you're closer than you think. The path forward is an evolution, not a rewrite. Read More...

• Erik Wilde [Head of Enterprise Strategy — Jentic]

The standup that broke me (and what I learned). We upgraded our stack. We adopted AI. We forgot to upgrade ourselves. 

Future of Engineering Stage — 30 minutes Future of Engineering Software engineering

Technology transformed overnight. Our team rituals and culture had to keep up. We were running 2026 engineering teams on 2005 ceremony software and wondering why transformation initiatives kept stalling halfway.  This talk is a love letter and a post-mortem. Starting from one painfully relatable standup that went wrong, we'll explore why the hardest part of modernizing engineering isn't the architecture — it's the unwritten rules, the unreasonable expectations, the inherited ceremonies, and the habits nobody ever put on a roadmap to change. It’s about how we make decisions faster, build trust differently, and design our human interactions with the same energy we bring to our systems. Read More...

• Iman Zarrabian [CTO — Ipsos Askia]

Why great engineers still struggle to deliver

Future of Engineering Stage — 30 minutes Software engineering Clean Code

This talk looks at a common belief in tech teams: that projects fail or are late because there are not enough good engineers. While strong engineering skills are important, they are often not the main reason delivery is slow or difficult.  The talk explains how delivery problems usually come from the environment around the team. This includes things like changing priorities, work that depends on other teams, slow feedback, and technical debt built over time when short‑term fixes are chosen instead of long‑term solutions. The session also introduces simple ideas from systems thinking, such as how company structure can affect the way systems are built and how fast teams can deliver. One example is Conway’s Law, which shows the link between how teams are organized and the software they create. The goal of the talk is to help people better understand what really impacts delivery, and how teams and organizations can create better conditions for success, not just better code.  Read More...

• Florentina Saizu [Senior Project Manager — MassMutual Romania]

Agent Led Data Engineering on Databricks: Build ETL pipelines, Train ML models & Run Inference

Future of Engineering Stage — 45 minutes

• Adarsh Nandan [Lead Scale Solution Engineer — Databricks]

International speaker keynote

DevOps Stage — 45 minutes

• Fatima Sarah Khalid [Senior Developer Advocate — GitLab]

Partner keynote

DevOps Stage — 30 minutes

Keynote powered by Metaminds

DevOps Stage — 30 minutes

Watching LLMs: observability best practices for AI systems

DevOps Stage — 45 minutes DevOps Open Source LLM Real case studies AI Systems

When you deploy an LLM-powered system, you might not always think about monitoring. But what if your RAG pipeline is not retrieving the right context, if your prompts are causing hallucinations, or if your token costs just tripled overnight? This talk covers the practical observability stack needed for LLM applications in production. By diving into code, we’ll explore how to instrument token usage and cost tracking across model calls, implement evaluation pipelines that catch quality degradation before your users do, and set up guardrails that prevent both prompt injection attacks and unexpected model behavior. You'll learn to trace requests through vector database queries, embedding generations, and completion calls to debug the full retrieval and generation pipeline.Using open source tools and vendor-agnostic approaches, we'll examine real production patterns: distinguishing between model latency and retrieval latency, tracking semantic drift in your vector embeddings, correlating user feedback with model parameters, and building dashboards that help you answer questions like "why did this query return irrelevant results?" and "which prompts are consuming 80% of our budget?"Whether you're running a chatbot, a document Q&A system, or any application with LLMs and vector databases, you'll leave with concrete strategies for understanding what your AI system is doing, how much it's costing you, and most importantly, whether it's working correctly.Takeaways:           •           How to implement observability across your entire LLM application           •           How to setup cost monitoring that prevents surprises           •           How to build effective guardrails and evaluation loops for production safety Read More...

• Gerard van Engelen [Founder — Obserfana]

Lunch Break

DevOps Stage — 1 hour

Software Delivery is about to get uncomfortable

DevOps Stage — 45 minutes DevOps Real case studies AI Agents

Software delivery is no longer just about faster pipelines. AI agents, platform engineering, policy-as-code, security automation, and developer experience are reshaping how software moves from idea to production. The next shift in DevOps is not only technical. It changes workflows, team responsibilities, and how engineering organizations make decisions.An atypical keynote with a lot of demos. Read More...

• Emanuele Bartolesi [GitHub Tech Lead — Xebia]

Faster, Leaner, Predictable: Quarkus vs Spring Under Real Load

DevOps Stage — 30 minutes DevOps Kubernetes Microservices

Modern organizations increasingly adopt FinOps principles to balance engineering autonomy with responsible cloud spending. In this presentation, I explore a practical comparison between Quarkus and Spring Boot through the lens of cost-efficient microservice design.  Using identical REST microservices deployed in containerized environments, I analyze how startup time, memory footprint, container size and runtime behavior directly influence infrastructure elasticity, compute utilization and long-term operational predictability.Rather than focusing solely on framework capabilities, the session highlights how technical decisions made during development cascade into real financial impact in production environments. Attendees will gain a clear understanding of when each framework is more cost-effective, how resource profiles affect Kubernetes scaling strategies, and how engineering teams can align architecture choices with measurable FinOps outcomes. This talk is designed for architects, developers and engineering leaders seeking a balanced, data-driven perspective on building faster, leaner and more predictable microservices. Read More...

• Alexandru-Cătălin Diniș [CTO — BMW Techworks Romania]

Modernizing Java with Confidence: GitHub Copilot and OpenRewrite in CI/CD Pipelines

DevOps Stage — 45 minutes Java DevOps Automation

Modernizing Java applications often means dealing with framework upgrades, large refactorings, dependency conflicts, and preparing the codebase for cloud-native deployments, all while keeping CI/CD pipelines stable and predictable.
In this session, I will show how DevOps and platform teams can combine AI-assisted tools such as GitHub Copilot with deterministic automation like OpenRewrite, integrated directly into GitHub Actions–based CI/CD pipelines. We will look at practical patterns for automating Java modernization safely, keeping control and reviewability of changes, and scaling these workflows across teams. Read More...

• Sandra Ahlgrimm [Senior Cloud Advocate — Microsoft]

Should This Be a Kubernetes Operator?

DevOps Stage — 30 minutes DevOps Kubernetes

A pragmatic look at what Kubernetes operators are, when building one makes sense, and how they enable opinionated resources in cloud-native environments.

• Alexandru Dejanu [Site Reliability Engineer — Alchemy]

The Rise of "Vibe Coding" & Its Impact On Software Security

Cybersecurity Stage — 45 minutes Security Software Development Vibe Coding

Software development is shifting from syntax-driven to intent-driven. This is transforming not only the software developer's workflow, but also who in an organization feels empowered to code in the first place. As application innovation skyrockets, how do we stay strongly grounded in security? This session discusses the security implications of the ever-growing "vibe coding" trend and proposes actionable solutions for leaders, developers, and hardcore vibe coders alike. Read More...

• Erika Dietrick [Software Engineer — Independent]

Partner keynote

Cybersecurity Stage — 30 minutes

Vulnerability Management Process Evolution: The Journey Behind It 

Cybersecurity Stage — 30 minutes

Vulnerability management is a field where everyone agrees on the theory - discover, report, triage, fix, verify, repeat - but how it's put into practice can look very different across organizations.  The interesting part isn't just the framework itself; it's the journey many teams take - starting from an email thread or a shared Excel file, evolving through Jira labels and a security-focused Slack channel, and eventually moving towards dedicated ASPM platforms and tools like DefectDojo.    Every step of that evolution is driven not by best-practice checklists, but by the business's own growing pains - its scale, its risk appetite, and the regulatory or client pressure it faces. In this talk, we'll walk through how vulnerability management processes mature in lockstep with an organization, what triggers each leap to the next level of complexity, and what the real cost of that leap looks like in terms of people, processes, and tooling. Most importantly, we'll ask the question nobody puts in the whitepaper: is it actually worth it?   Read More...

• Dmytro Tereshchenko  [Chief Information Security Officer — Sigma Software Group]

Me, Myself & AI: Compliance, Control and Trust in an Algorithmic Age

Cybersecurity Stage — 45 minutes Artificial Intelligence Machine Learning Security Real case studies

AI and machine learning promise transformative benefits, from efficiency and innovation to enhanced cyber defense. Yet this same technology is a double-edged sword. The rapid adoption of AI introduces new compliance challenges around data privacy, bias, transparency, third-party risk, and ethical use—often at a pace faster than organizations can govern. As automation enables cybercriminals to launch attacks at unprecedented scale, it simultaneously becomes one of our most powerful defensive tools. The question is no longer whether to use AI, but how to do so responsibly.This talk explores how emerging global regulations such as the EU AI Act, U.S. legislative efforts, NIS2 and DORA are reshaping accountability for organizations that create, deploy, or rely on AI systems—particularly through third parties. Taking a risk-based approach, we examine what “high-risk AI” really means, why regulatory clarity is essential to avoid stifling innovation, and how standards such as ISO/IEC 42001 (AI Governance), 23894 (AI Risk Management), and 27090 (AI Security Controls) can provide practical guardrails. With data sovereignty becoming a battleground, geopolitical tensions fragmenting technology ecosystems, and over half of security incidents now linked to third parties, compliance is converging into a single concept: Digital Trust.Attendees will leave with a clearer understanding of how to integrate AI into compliance frameworks without losing control—knowing where their data is, who can access it, and which laws apply—while balancing innovation, security, and societal values. Because in an AI-driven world, trust isn’t a by-product of technology; it’s a strategic choice. Read More...

• Mathew Caplan [Director of International Professional Services — Orange Cyberdefense]

Lunch Break

Cybersecurity Stage — 1 hour

When the Attack Landscape Is Re‑Mapping Every Minute, How Can the Defense Keep Pace?

Cybersecurity Stage — 45 minutes Security

Imagine a world where the threat landscape shifts every second. Vulnerabilities emerge and disappear before you can document them. Attack infrastructure rotates faster than you can block it. Threat actors deploy new tactics while you’re still analyzing the last one. You are right, there’s no need to imagine it, that’s exactly the world we live in.The velocity is relentless. The gap between attack and defense widens by the hour. The traditional playbook, perimeter defense, static threat models, reactive response cycles, was built for a different landscape entirely.The reality is not just the scale of attacks, but the velocity at which the attack surface transforms. New entry points proliferate, infrastructure rotates, tactics shift. All while defenders react to yesterday’s threat. Traditional boundaries can no longer contain this.We will be addressing other, maybe harder, and maybe more nuanced questions. Are we measuring the right things? Are we defending the right assets? Do we understand where our vulnerabilities and weaknesses lie?Third parties and supply chain partners are more than ever part of our attack surface. They’re embedded in the risk landscape, and that changes what need to be protected.In an environment where threats continuously evolve, your defenses are only as strong as your extended network. Organizations that ignore third-party and supply chain risk will find their defenses incomplete.Agility today isn’t just about moving fast or being flexible, it is about asking the right question quickly enough to stay ahead of the threat. Read More...

• Andreea Gulie [Territory CISO — BNP Paribas]

When Bots Think for You: AI Automation and the New Frontiers of Data Privacy

Cybersecurity Stage — 30 minutes Artificial Intelligence Real case studies Automation

As artificial intelligence increasingly automates complex cognitive tasks, organizations face the challenge of balancing technological efficiency with the protection of personal data. Intelligent systems now collect, analyze, and act on vast amounts of user information, reshaping traditional ways of working.  This keynote explores how AI‑driven automation is transforming everyday operations, drawing on practical case studies and recent research to showcase innovative approaches that boost productivity and streamline workflows, while introducing a framework for responsible automation, emphasizing data minimization, ethical design, and regulatory alignment - demonstrating how organizations can harness AI’s potential while safeguarding individual rights and maintaining trust. Read More...

• Anca Grigoras [Data Privacy Services Team Lead — Allianz Services Romania]
• Evelin Bratu [Global Service Owner — Allianz Services Romania]

AI and the Cloud: Securing the Uncontrolled

Cybersecurity Stage — 30 minutes Artificial Intelligence Real case studies

In an era of managed services and AI pipelines where control is delegated to third party services, security comes in as a main responsibility for the developers. What is riskier?  A loose firewall in the VPC or a non guard railed AI model handling your data processing? Is your AI deleting your production database while attackers explore your instances like a museum without the “No touch” signs? Explore new ways of protecting your business using examples from VOIS projects: Cavendish and AI Booster Responsible AI Layer. Read More...

• Alexandru Gheara [Site Reliability Engineer — VOIS]
• Iustin Dinca [Software Engineer — VOIS]

International speaker keynote

Cybersecurity Stage — 45 minutes

Staying Relevant in QA: Building Skills, Communities, and Strategies for the Future

Quality engineering & Automation Stage — 45 minutes QA Community Impact Skill Growth

In today's fast-paced work environment, QA professionals face the challenge of staying relevant—whether as testers, consultants, entrepreneurs, or team managers.  This keynote explores practical strategies for continuous competence development, fostering learning cultures, and leveraging communities to thrive in evolving landscapes. We’ll discuss organisational frameworks like DORA, the role of training and certifications, and how to build resilient teams that embrace innovation. Read More...

• Mikko Paloheimo [QA Coach — Paloheimo QA Mentorworks]

Continuous Quality Engineering: Building Confidence Across the SDLC

Quality engineering & Automation Stage — 30 minutes

Ever shipped a feature that looked flawless in staging, only to see it break in production?  You’re not the only one. Today, the real bottleneck isn’t how fast we can ship—it’s how confident we feel about what we ship. Teams now have the tools to automate tests, build pipelines, and deploy at lightning speed, but quality issues still surface too late—sometimes right after a release, sometimes after users have already been impacted. Continuous Quality Engineering flips that script. Instead of treating quality as a final gate, we treat it as a continuous feedback system that spans the whole SDLC: from requirements and development, through CI/CD validation, performance engineering, observability, and production monitoring. In this session we’ll make the concept practical by focusing on three interconnected practices: ·       Shift‑left validation – moving testing and verification earlier so defects are caught when they’re cheapest to fix.·       CI/CD quality gates – embedding automated checks directly into the pipeline to enforce standards on every commit. ·       Shift‑right observability – keeping an eye on what happens in production, turning real‑world data into actionable insights for the next iteration. The aim isn’t merely to increase test coverage; it’s to create meaningful feedback loops to help teams detect risks earlier, make better release decisions, and continuously improve the quality of their systems and applications. “Let’s stop treating QA as a last‑minute checkpoint and make it a continuous companion that walks with every commit, every build, and every release.”  Read More...

• Daniel Fifiita [Performance Testing Tech Lead — BNP Paribas]

The Signals We Keep: Friction, Feedback, and Appropriateness in Delivery Systems

Quality engineering & Automation Stage — 45 minutes Testing Q&A Systems Automation

The world is noisy. Overwhelmingly so at times.Engineering teams have spent years trying to quieten that noise. We remove friction, smooth developer experience, and optimise delivery systems to feel faster and cleaner. But lately I’ve been wondering whether we’ve been quietening the wrong signals.Much of this comes from real frustrations in delivery systems where confidence is often inferred from outputs, dashboards, and automation. In testing, we’ve long understood the importance of storytelling and appropriateness in how we interpret those signals, yet even in highly optimised environments, teams can still struggle to understand how their systems actually feel to work within, and what feedback they can truly rely on.As the actors in software delivery evolve, and the systems around us change faster than the expectations we built them on, some of the feedback that once helped us feel oriented has started to disappear. Pipelines are polished, tooling is optimised, and yet that sense of orientation is often harder to find.In this talk, I explore friction through a feedback lens. Drawing on personal frustrations, changing industry expectations, and lessons from real delivery environments, I’ll look at how friction can act as a signal rather than a destination. Not something to eliminate by default, but something that helps us notice when systems are renegotiating their identity and balance.This is a moment to reconsider how delivery systems feel, what feedback still matters, and how we decide which signals are appropriate to keep as our delivery engines evolve.Takeaways:1. Use a simple loop: Notice → Interpret → OrientWhen something feels off, treat it as a signal, not a problem to remove.2. Interpret signals empiricallyTreat outputs (AI, dashboards, automation) as evidence, not truth, and make assumptions explicit.3. Create moments to reorientUse conversations and shared understanding to sense-check what your systems are telling you before acting. Read More...

• Chris Armstrong [DevRel Test Advocate — SmartBear]

Lunch Break

Quality engineering & Automation Stage — 1 hour

User Experience Is Part of Quality — But Who Tests It Before the Customer?

Quality engineering & Automation Stage — 45 minutes Testing Real case studies

Modern software teams are great at testing functionality but user experience often slips through the cracks. In this talk, I explore a simple but critical question: who actually tests the experience before the customer does?  Through real world examples from product development, we’ll look at situations where everything “worked” yet the we fall. From misunderstood flows to broken assumptions and trust issues, these cases reveal a common gap in how we think about quality.
I’ll share practical ways to bring user experience into everyday testing: from early design reviews and collaborative testing sessions to validating assumptions before release. Because if we don’t test the experience, our users will. Read More...

• Pekka Pönkänen [Engineering Team Lead — Wolt]

QA in the Age of AI: From Testing to Trust

Quality engineering & Automation Stage — 30 minutes Artificial Intelligence QA Testing Real case studies

In a world where code can be generated in seconds with the help of AI, testing doesn’t disappear — it becomes more critical than ever. But what does “quality” really mean when delivery speed is growing exponentially and applications are becoming increasingly complex? In this talk, I’ll explore the evolution of software testing from the perspective of an engineer who moved from manual testing to automation with Selenium, and more recently to modern testing using Playwright and TypeScript. We’ll discuss how the QA role has evolved within CI/CD pipelines and how AI tools like ChatGPT and GitHub Copilot are reshaping the way we write and validate software.Through real examples and a live demo, I’ll show how AI can accelerate development, but also how it can introduce subtle issues that only a well-designed testing strategy can uncover. We’ll also look at why “more tests” doesn’t necessarily mean “better quality.” Read More...

• Cosmin Trandafir [IT Systems Testing Consultant — Bug Hunter]

How AI helps in digital content production: scaling, automation and creativity at 100x

Quality engineering & Automation Stage — 30 minutes Testing Real case studies AI Systems

The digital content landscape has shifted from fast-paced to exponential, leaving traditional workflows at a breaking point.  This session explores how to re-architect the creative process by moving beyond traditional productions and basic prompts into fully integrated AI ecosystems that enable a 100x leap in output. We will dive into the practicalities of automated multi-channel distribution, the evolution of the creator’s role from "maker" to "director," and the tools that finally break the trade-off between quality and quantity. Whether you are a solo creator or a global brand, you’ll walk away with a blueprint for eliminating production bottlenecks and scaling your creative vision at the speed of thought. Read More...

• Dr. Péter Szántó [Co-Founder & CPO — reel8 Digital & AI Production House]

Crafting User-Friendly Test Automation

Quality engineering & Automation Stage — 45 minutes

Crafting a User Friendly Test Automation Framework : Yes, you got it right.  We will talk about the user-friendliness of not the product we are developing, but our test automation framework. Well, the test automation framework is a product indeed, right? Let’s see ways to build a user-friendly test automation framework. Read More...

• Mesut Durukal [QA Capabilities Leader — OMP]

Hugging Face ecosystem

AI Product Stage — 45 minutes Machine Learning AI Systems

Niels will showcase the Hugging Face ecosystem, all the tools and services Hugging Face provides for making machine learning accessible, open and collaborative.

• Niels Rogge [Machine Learning Engineer — Hugging Face and ML6]

From Chatbot to Multi-Agent AI: ELLA's Journey to Operational Excellence

AI Product Stage — 30 minutes AI Systems AI Agents

ELLA began as a simple AI chatbot for answering operational questions. Today, it's evolving into a sophisticated multi-agent AI architecture that's transforming how we approach operational excellence at scale.  This session chronicles the program management journey of transforming ELLA from a single-purpose conversational tool into an intelligent, modular multi-agent system. We'll explore how we orchestrated this evolution across five cross-functional teams (Data Engineering, Data Science, Software Engineering, Support Engineering, and Data Analysis) while maintaining operational continuity. Key topics include: architecting for extensibility with pluggable agents, managing the technical transition from monolithic to distributed AI systems, balancing innovation velocity with operational stability, and measuring success beyond traditional metrics. Attendees will learn practical frameworks for managing AI transformation programs, strategies for coordinating multi-disciplinary teams on emerging technology initiatives, and approaches to building scalable AI solutions that grow with organizational needs. This is a story of program management excellence applied to cutting-edge AI - turning ambitious vision into operational reality.  Read More...

• Marius Cristea [Sr. Technical Program Manager — Amazon]

Partner keynote

AI Product Stage — 30 minutes

Manipulating The Machine: Prompt Injections And Counter Measures.

AI Product Stage — 45 minutes LLM AI Systems

In the rapidly evolving landscape of Large Language Models (LLMs), GPTs, and the omnipresence of OpenAI, prompt injections have become a subtle but significant vector for causing harm to AI-based tools.  With seemingly simple yet creative techniques, it is often possible to extract information that was meant to remain hidden.We will explore various methods to perform prompt injection to extract system prompts and documents used by GPTs and other LLM-based tools. Along the way, we will discuss ways to integrate countermeasures into these tools to defend against attempts to steal information.Learn about prompt engineering, prompt injection, protective measures, and why, after all, we should not worry too much. Read More...

• Georg Dresler [Principal Software Architect — Ray Sono]

Lunch Break

AI Product Stage — 1 hour

Keynote powered by Tremend - Publicis Sapient

AI Product Stage — 30 minutes

Real time visualizations with Fabric?

AI Product Stage — 30 minutes Artificial Intelligence Product IoT

The most common IoT scenario is getting data on some dashboard. Power Bi to the rescue, everybody is happy. But wait Real-time streaming in Power BI is on way to retirement. What to do? Another superhero jumps in Microsoft Fabric Real-Time Intelligence, and can do the job, let’s see how. And wait, Real-Time Intelligence is much more than a replacement for Real-time streaming.So, let’s explore Fabric Real-Time Intelligence, what it is, what can do and what to do exploit to the max.Of course, it will be a sweet session, Kinder chocolates sweet. Read More...

• Catalin Gheorghiu [Solution Architect II — EPAM Romania]

Partner keynote

AI Product Stage — 45 minutes

Accelerating Innovation with AgenticAI

AI Product Stage — 45 minutes Software Development Agentic AI

Agentic AI is the most marketable buzzword in 2026. However, does it really improve your business? Accelerating only a part of your Software Development Lifecyle is useless. It's like using a Ferrari while you are still driving on a muddy road. If you truly want to accelerate, you need access to a highway with fast access lanes, efficient gateways, and secure guardrails. That's where a full DevSecOps platform, powered by AI, makes sense: it provides you with everything you need to accelerate the whole system and lifecycle, equipping everyone in your organization with the capabilities they need to perform fast, efficiently, and securely. Join this session to see how GitLab Duo Agent Platform can help you going faster, wherever you are: in the GitLab GUI or in your IDE. You will see how to handle your backlog faster, transform it into working code, properly tested, and resolve vulnerabilities while having you full control over your agents to stay within your corporate boundaries and mandatory compliance frameworks. Read More...

• Olivier Dupré [Solutions Architect — GitLab]

An Agent Walks Into a Bar... and Meets Another Agent: Multi-Agent Systems in Action

AI Stage — 45 minutes AI Systems AI Agents

In the world of AI, a single agent can accomplish a lot—but what happens when multiple agents enter the scene? This session is your ultimate guide to understanding multi-agent systems (MAS), the key to building scalable, intelligent, and collaborative AI solutions. We'll break down the core distinctions between single-agent and multi-agent architectures, explore real-world applications, and dive into the complexities of agent interaction, coordination, and fault tolerance. Whether you're designing autonomous trading bots, smart logistics networks, or AI-driven simulations, this talk will give you a solid foundation in MAS and its transformative potential. Get ready for a deep dive into the tech that’s reshaping AI collaboration!- Why one “super agent” quickly gets overwhelmed- How specialized agents working together outperform a jack-of-all-trades- The five orchestration patterns (Sequential, Concurrent, Group Chat, Handoff, Magnetic) that make multi-agent systems effective in real-world use cases- Where this is heading: emergent behaviors, cross-domain teams, and self-healing systems- How to start building your multi-agent systems Read More...

• Maxim Salnikov [AI Dev Tools & Platforms Solution Engineer — Microsoft]

From API to AI: A Survival Guide for Building an MCP Server

AI Stage — 30 minutes Artificial Intelligence LLM APIs

We started with a successful B2B API, but quickly realized that feeding it directly to an LLM serving human travelers required a translation layer.  This talk explores how we built a consumer-facing MCP server that bridges that gap — translating rigid API schemas into AI-friendly tools that understand "budgets" and "vibes," deploying a single codebase across multiple AI partner integrations with different auth models and platform constraints, and ensuring reliability through an LLM evaluation framework with "golden prompts," data minimization, and prompt injection defenses.  Read More...

• Hakan Turkmenoglu [Full Stack Software Engineer — Booking.com]
• Rail Mazgutov [Backend Engineer — Booking.com]

AI is Now Everyone's Problem - Act Like It!

AI Stage — 30 minutes Artificial Intelligence Generative AI

There was a time when "AI in the enterprise" meant a data science team, a Jupyter notebook, and a model that ran once a day in a corner of your data warehouse. Controlled, Contained, rarely customer-facing. That era is over. GenAI didn't just raise the ambition for what AI can do, it changed who is responsible for delivering it. Enterprise developers are now building AI-native applications that are end-user facing, always-on, and deeply integrated with business-critical workflows. We went from deterministic logic to probabilistic models, from scripts to autonomous agents, and "it works in the sandbox" is no longer good enough.Data scientists haven't disappeared — but their role has evolved. Fine-tuning models, curating training data, and evaluating quality are more critical than ever. The difference is that their work now feeds directly into production systems built and operated by enterprise developers. Two disciplines, one pipeline, and the handoff between them needs to be seamless.This keynote tackles the hard questions that come with AI growing up: How do you move from prototype to production without losing control? How do you build with confidence while keeping security, compliance, and governance intact? And how do you bridge the gap between experimentation and enterprise scale?If your team is done experimenting and ready to ship, this talk is for you. Read More...

• Timo Salm [Master Solutions Architect — Broadcom]

Architecting Data for AI Agency

AI Stage — 45 minutes

As Agentic AI introduces unprecedented, dynamic traffic patterns, data teams must rapidly rearchitect and optimize their infrastructure. In this session, we will explore the evolution of data preparation and cleaning for the Agentic AI workstream and the broader ecosystem. These new workloads are required to support new demanding business requirements. Crucially, we will unpack how robust governance, specifically discoverability, enriched metadata, and deterministic lineage, has become the operational engine enabling agents to autonomously reason and act. How to transform bottlenecks into optimized and scalable AI foundations.  Read More...

• Horia Niculescu [Data Analytics Specialist — Google]

Lunch Break

AI Stage — 1 hour

International speaker keynote

AI Stage — 45 minutes

• Paul Ardeleanu [Principal Architect (Developer Relations) — Vonage]

Spec-Driven Development with Kiro: Building AI-Aligned Software Through Living Specifications

AI Stage — 30 minutes

• Haim Michael [Software Developer — Trainer | Consultant]

AI for Delivery Teams

AI Stage — 30 minutes Artificial Intelligence Software engineering

The most visible use of AI in engineering teams right now involves coding agents, and the attention they get is well deserved. They help engineers write code faster, catch bugs they would otherwise miss, expand their debugging capabilities and reduce boilerplate.  But when delivery teams consider their next steps with AI, they tend to move phase by phase through the SDLC (requirements, design, build, test, deploy, monitor) and try to plug AI into each one. However, delivery teams don't run on code alone, nor are they composed only of engineers. They run on decisions, specifications, and context that someone must curate and track. That work crosses every phase and belongs to no phase. And when it breaks down, it doesn't show up as a bug, but as a week lost onboarding someone, a spec that contradicts itself or the code, or a decision nobody can trace back to. This keynote argues that AI's role in software delivery extends far beyond writing code, and achieving this requires teams to change how they work, not just their tools.  Read More...

• Mihaela Maciuca [Senior Engineering Manager — Cognizant]

The Self-Healing Canary: Integrating Agentic AIOps into Your Releases

AI Stage — 45 minutes Artificial Intelligence Kubernetes AI Systems

Standard progressive delivery is an essential safety net, but in most cases, a rollback is just the beginning of a long investigation. This session explores how to evolve from simple automated rollbacks to a self-healing release cycle by integrating an agentic AI system via the A2A protocol.
We will demonstrate an architecture where specialized agents analyze canary logs and metrics in parallel to provide high-fidelity health checks. If an issue is detected, the system immediately signals a rollback while simultaneously launching an asynchronous remediation workflow. This "coding agent" analyzes the specific error against the app's source code to either create a detailed issue description or propose a fix in the form of a pull/merge request.
To maintain performance and keep costs low, the system utilizes a hybrid model strategy. We run Small Language Models (SLMs) directly on a Kubernetes cluster for fast, local analysis and escalate to larger frontier models only when complex reasoning is required. Attendees will learn how to build this modular system and get access to the open-source implementation to start automating their own remediation loops. Read More...

• Kevin Dubois [Senior Principal Developer Advocate — IBM]