Everyone can build: empowering non-engineers to contribute to the code

Programming Stage — 45 minutes

While the industry keeps debating whether AI makes engineers a bit more productive or a lot more productive, the debate misses the real shift.  What's happening is a collapse of the talent stack: designers, PMs, ops… suddenly able to contribute directly to the codebase. And that changes the mandate for platform teams. It's no longer just about enabling engineers to go faster, but enabling everyone to build safely, responsibly, and with as little friction as possible.At Alan, a health insurance company serving over 1 million users in production across Europe, we embraced this shift head-on by launching "Everyone Can Build." In just two quarters, 283 pull requests from non-engineers were merged to production—from designers shipping UI improvements to operations teams enhancing internal tools. This isn't a proof of concept: it's a production-grade transformation powered by AI that has fundamentally redefined our development culture and what it means to be a platform engineering team.This session reveals the blueprint of how we achieved 100% AI adoption across our organization and rebuilt our engineering platform to support this new reality. Read More...

• Sofia Lescano [Senior Software Developer and Team lead — Alan]

How to Build an AI System That Makes It to Production

Programming Stage — 30 minutes

Demos are easy. Production GenAI systems are not. Drawing on lessons from three GenAI products shipped to production, this talk explains the core concepts behind building reliable GenAI systems and the architectural decisions that matter most in practice. We’ll look at where the model sits in the system, what responsibilities it should own, how it should use tools and workflows, and how the context provided to the LLM - user information, task data, business knowledge, memory, and system state - can determine the quality of the final result.The goal is to show how to move from a promising prototype to an AI system that reaches production and stays there. Read More...

• Florian Minel Ulmeanu [Senior AI Engineer — IsoSkills]

Building AI-powered products in the Agentic Age

Programming Stage — 45 minutes

Artificial intelligence is entering a new phase, bringing new capabilities such as reasoning and planning into products and services. In this talk, we will explore how to build AI-powered products and discuss the key considerations for creating experiences that are intelligent, useful, and reliable at scale.

• Silviu-Tudor Serban [Staff Engineer, AI — TCN]

Unmasking Traefik: Multi-Domain Routing Patterns in Kubernetes

Programming Stage — 30 minutes

Ever wondered how Traefik manages domain magic in the background? We’ll solve the mystery of having multiple domains in one environment, how it can improve cloud platform migrations, and create a seamless user experience for end-users.  Context: Traefik Proxy as an Ingress Controller in Kubernetes. Based on a real-world migration experience, this case study will walk you through the process of configuring not one, but three domains with a single IP!  It also improved our overall infrastructure in multiple ways, reducing resources being only one of them. Curious about another scenario? Find out how two domains can happily live together in a cluster, but be completely independent, each having their separate IP and microservices. This session is perfect for anyone curious to dive into Traefik’s capabilities and discover different uses of multi-domain setups. Code examples will cover two alternatives: Ingress and Gateway API. Read More...

• Edith Karda [Senior DevOps Engineer — .msg Romania]

Lunch break

Programming Stage — 1 hour

From Legacy Code to Cloud-Ready Applications with GitHub Copilot

Programming Stage — 45 minutes

Legacy modernization is slow, risky, and often difficult to scale.  In this session, we will see how GitHub Copilot Modernization Agent can help teams understand an existing Java or .NET application, create a practical migration plan, update the code and dependencies, and prepare it for Azure. We will also discuss what AI can accelerate—and what still requires experienced engineers. Read More...

• Radu Vunvulea [Group Head of Cloud — Endava]

Governing AI at Scale: How We Control and Optimize Azure & OpenAI with an Internal Developer Portal

Programming Stage — 30 minutes

As AI adoption grows, managing visibility, governance, and cost becomes increasingly complex. In this session, we’ll share how we use our Internal Developer Portal, to control AI resources across Azure, OpenAI, and AI Foundry. You’ll learn how we apply governance rules, and custom reporting to enforce security, ensure compliance, and gain full visibility—while also introducing cost insights bycorrelating token usage with pricing. A practical look at how to scale AI responsibly, balancing innovation with control. Read More...

• Ciprian Pop [Senior Systems Engineer — EPAM Romania]

Demystifying Workflows in Microsoft Agent Framework for .NET: Controlling Multi-Agent Chaos

Programming Stage — 45 minutes

Microsoft Agent Framework ships high-level orchestration patterns such as Sequential, Concurrent, Handoff, Group Chat, and Magentic, but real systems eventually need conditional branching, dynamic fan-out, or fine-grained control over what one agent hands to the next that goes beyond what those patterns expose. In this session, we show how edges and executors let us build carefully tuned workflows when the built-in patterns don't fit, and how those workflows can stay fully observable from end to end. We also compare when to stay at the high level and when to drop down to the workflow graph, using practical .NET examples where deterministic flow, clear state transitions, and agent coordination matter. The result is a mental model for building flexible agent systems without losing predictability or full observability. Read More...

• Daniel Costea [Senior AI Software Engineer — Author | Speaker | Trainer]

Engineering Decisions, Not Just Systems. Lessons from Cybersecurity, Systems Thinking and GRC

Programming Stage — 30 minutes

Modern engineering teams invest heavily in building resilient systems, yet many of the most impactful cybersecurity incidents are not caused by technology alone. They are the result of decisions, unclear ownership, organizational complexity, and failures in governance. This session explores cybersecurity from a different perspective: organizations as engineered systems. 
Drawing on real world incidents, we will look beyond technical root causes to uncover the organizational patterns that lead to major failures. Along the way, we will connect concepts from systems thinking, software engineering and Governance, Risk & Compliance (GRC) to show why resilient organizations require more than secure code. They require well engineered decisions. 
Whether you are a software engineer, security professional, architect, or engineering leader, you will leave with a mental model for understanding how technical excellence and organizational design work together to build resilient systems.  Read More...

• Alex Iacob [Information Risk Consultant — Mass Mutual Romania]

Architecture and Business in the AI Era

Programming Stage — 30 minutes

Evolutionary Architecture: Evolving Your Java Applications Through the Art of Decision-Making

Java Stage — 45 minutes

Join me for an engaging session that explores the indispensable role of decision-making in maintaining the evolvability of Java applications.  Drawing inspiration from Antoni Gaudí's visionary approach to crafting La Sagrada Família, we delve into the art of decision-making, highlighting its critical role in ensuring the longevity and adaptability of our systems.By the end of this talk, you'll gain a profound understanding of the significance of decision-making in shaping robust architectures and walk away with a practical toolkit of best practices—inspired by Gaudí's methodology—to effectively manage and document your architectural decisions.Don't miss this opportunity to elevate your architectural decision-making skills and fortify the foundations of your projects for the future. Read More...

• José Enrique Calderón Sanz [Senior Lead Engineer — JPMorganChase]

Modular Monoliths with Spring Modulith

Java Stage — 30 minutes

Microservices promise loose coupling, but you pay for it up front in network calls, distributed transactions, and the operational weight of many deployments. Microservices promise loose coupling, but you pay for it up front in network calls, distributed transactions, and the operational weight of many deployments.Most teams don't need that right away. They need modularization instead, and they need to avoid creating a big ball of mud in the long run.

In this talk, I'll walk through a running application to show how modules stay decoupled by communicating through Spring events instead of direct calls, how the Event Publication Registry gives you a built-in transactional outbox so a failed consumer never costs you data, and how module documentation and runtime observability come for free. Read More...

• Vlad Mihalcea [Java Champion — Moderator Java Stage]

If It Ain't Broke: The Hidden Cost of Your Java Comfort Zone

Java Stage — 45 minutes

Many Java teams treat their runtime version as a settled question. It isn't.  Every major Java release since 8 has delivered measurable performance improvements — in garbage collection, JIT compilation, memory efficiency, and concurrency — and staying behind means paying for that choice in CPU time, memory bills, and operational complexity.Before exploring what you're missing, we'll clear up a widely misunderstood clue hiding in your build tool: the difference between the Java version your code is written against and the JVM version it actually runs on — a distinction with real consequences that most teams discover at the wrong moment.From there, we trace a decade of JVM evolution: from G1 to ZGC and Shenandoah, JIT advances, compact strings, and Project Loom's virtual threads — each backed by concrete numbers.The case is simple: staying on an old Java version isn't caution. It's a tax. Read More...

• Konrad Szałkowski [Principial Software Engineer — Egnyte]

The happy path is a lie: Modeling failure end-to-end

Java Stage — 30 minutes

"Something went wrong." We’ve all shipped that screen, it's what's left when the happy path was the only path we built. This talk is about making it impossible.  We build the happy path, it works under normal conditions, it demos well, it ships, and then we get judged on everything it ignored: the payment that fails, the session that expires, the two people editing the same database record. Those failure paths are where the incidents, the tickets, and the bad reviews actually live, and most of us leave them to chance: a thrown exception here, a generic catch there, a spinner that never resolves. 
Not this time. We'll treat failures as first-class, typed values we can reason about, not exceptions thrown into the dark, and follow a single failing action through the whole system. The hard part is the bit almost none of us does, keeping that guarantee alive across the wire, where a domain error turns into JSON and lands on the other side as a generic "Error" and a loading state. We'll carry it across intact, from the domain model, through serialisation, to the exact thing the user sees. 
By the end, "Something went wrong" won't be a screen we ship, it'll be a bug that won't compile. Examples are in Kotlin and TypeScript, but the ideas land in any language growing sealed types, records, and pattern matching –modern Java included. No functional-programming background required; bring your worst error screen.  Read More...

• Gabriel Bornea [Chapter Lead Engineer — ING Hubs Romania]

Lunch break

Java Stage — 1 hour

From AI to Agent with Langchain4J

Java Stage — 45 minutes

You’ve built AI features into your application. Models are wrapped in services, RAG is in place, tools are wired, and calls are flowing. Then requirements evolve. A single response is no longer enough. You need steps that follow each other, branches based on decisions, retries when things fail, and sometimes multiple actions happening at the same time. At that point, it becomes unclear how to structure the logic without losing control or readability.
In this talk, we move from AI calls to agentic systems using LangChain4j. We’ll explore common workflow patterns such as sequential, loop, parallel, and conditional flows, along with error handling as a first-class concern. From there, we’ll zoom out to agentic concepts like shared state with an AgenticScope, goal-oriented agents, and pure agentic setups. We’ll also look beyond AI-only approaches, mixing in non-AI agents and human-in-the-loop interactions. All of this will be built live, step by step, to show how these ideas work in practice.
You’ll leave with a concrete understanding of when to use workflows and when to rely on agents, how to combine both without confusion, and how to keep state, goals, and responsibilities explicit. The goal is not to build smarter prompts, but systems that are easier to reason about, easier to evolve, and better aligned with real-world application needs. Read More...

• Loïc Magnette [Senior Java Developer — Oniryx]

The AI-Native Developer: Learning New Technology Stacks Faster Than Ever

Java Stage — 30 minutes

A few years ago, switching technology stacks often meant weeks of training, documentation reading, and mentorship. Today, AI coding companions are reshaping that journey.  Using a likely scenario of Java engineers onboarding to a .NET microservices project, we'll show how GenAI can accelerate architecture discovery, code comprehension, framework learning, and day-to-day development activities. We'll discuss both the opportunities and the limitations, highlighting how developers are evolving from memorising frameworks to orchestrating knowledge, validating AI outputs, and focusing on higher-level engineering decisions. The session is not about replacing developers with AI. It's about how developers can leverage AI to become effective in unfamiliar environments faster than ever before.  Read More...

• Silviu Cucuiet [Software Engineer — Cognizant Romania]
• Adriel Iclodean [Senior Software Engineer — Cognizant Romania]

Accelerating Maven Builds: From Snail's Pace 🐌 to Rocket Speed 🚀

Java Stage — 45 minutes

Are you tired of watching Maven builds crawl at a snail’s pace, wasting precious development time? Spending too much time at the coffee machine, or fighting battles with wooden swords under the excuse “my code’s compiling”? Join me to learn how to supercharge your Maven build! I’ll cover three main steps to start speeding up your project build. Learn how each speeds up your build, when they provide the biggest gains, and what pitfalls await.Take the next step in boosting your developer productivity by learning practical tips to decrease context switching and increase development speed and feedback cycle. Your journey from snail’s pace to rocket speed begins today! Read More...

• Maarten Mulders [Solution Architect | Trainer | Technology Advocate — Info Support]

Standing Still Is the Real Risk: Java, AI and the Cost of Comfort

Java Stage — 30 minutes

Java isn't standing still, and neither is the role of the senior engineer building on it. In this closing panel, four international experts sit down to talk evolutionary architecture, breaking out of technical comfort zones, building AI agents on classic Java foundations, and the unglamorous engineering work that quietly makes or breaks a team's efficiency. The conversation then opens up into the topic on every senior specialist's mind: how is the job actually changing, what's worth learning next, and how do you stay relevant in a market being reshaped by AI. Without the hype and without the panic.As the last session of the day, we've left plenty of room for audience Q&A, so bring your questions.

• Konrad Szałkowski [Principial Software Engineer — Egnyte]
• Loïc Magnette [Senior Java Developer — Oniryx]
• Maarten Mulders [Solution Architect | Trainer | Technology Advocate — Info Support]
• José Enrique Calderón Sanz [Senior Lead Engineer — JPMorganChase]

Learning Agent Harnesses with Robby’s Iterative Evolution in Microsoft Agent Framework for .NET

Roundtable Area — 1 hour

A chatbot does not become an agent system by adding every capability at once. It evolves when real failures reveal what is missing. Using Microsoft Agent Framework in .NET, follow Robby, a robot car, as he develops from a simple chatbot into an agent system with memory, context, tools, middleware, evaluations, and workflows. Each stage begins with a limitation Robby cannot handle, and introduces only the harness capability needed to solve it.You will also examine sequential, concurrent, and handoff orchestration, then transition to explicit workflows for cases that require predictable control.With practical C# examples, learn when to rely on agent reasoning, when to encode rules in deterministic code, and why robust agent systems should grow from observed problems rather than premature complexity. Read More...

• Daniel Costea [Senior AI Software Engineer — Author | Speaker | Trainer]

Lunch Break

Roundtable Area — 1 hour

When AI Can Do Almost Anything, What Do You Bring to the Table?

Roundtable Area — 1 hour

AI can already write code, analyze data, create content, conduct research, and build prototypes in minutes. As execution becomes faster and more accessible to almost anyone, an important question emerges: what makes our work valuable?In this talk, we go beyond prompts and explore the skills and tools that become increasingly important as AI becomes capable of doing more.We’ll look at critical thinking, creativity, expertise, and human judgment and how AI can amplify these abilities.Because when AI can produce more, faster, the difference lies with the person who knows what is worth building, what is worth keeping, and which decisions are worth making. Read More...

• Alin Chiriac [Founder and CEO — Adfaber]

Evolutionary Architecture in Practice: Making Better Decisions as Your Java Application Evolves

Roundtable Area — 1 hour

The idea would be to take the concepts from the talk a step further and make the session much more hands-on and interactive. Rather than focusing on finding one “correct” architecture, the goal would be for participants to practise evaluating trade-offs, understanding the consequences of their decisions, and evolving the architecture incrementally. Read More...

• José Enrique Calderón Sanz [Senior Lead Engineer — JPMorganChase]

The AI-First SDLC: Reimagining Software Engineering

Roundtable Area — 1 hour

Coding is no longer the bottleneck - context and alignment are.  This talk presents the AI-first SDLC we’ve adopted across our engineering teams and client projects, built on three pillars: durable project context, living specifications as the single source of truth, and chained agentic loops that implement and review against those specifications. We’ll cover how the model works, how to bootstrap it in an existing codebase, and what we learned from putting it into practice. Read More...

• Evgeni Rusev [Solution Architect — Tecknoworks]