A short, honest morning session where developers share their real-world fuckup stories — what went wrong, what they learned, and how these moments shaped them into better engineers. A candid, funny, and insightful start to the day.

AI is the reason enterprises are rethinking their legacy systems - and the tool that makes this transformation possible. This talk examines how the push to stay relevant in the AI era meets the power of AI itself to drive modernization and unlock business value.

‍

New LLM-based projects bring new headaches — from context limits and prompt misuse to rising costs and complex architectures. This talk explores real-world challenges and solutions: testing and validation, RAG vs. plugins trade-offs, multi-agent fact-checking, and securing your GenAI apps against data leaks and RAG poisoning.

Your startup won’t die from one bad decision — it’ll die from a hundred tiny ones. The skipped 1:1s, the tech debt you promised to fix “next sprint,” the ignored edge case that wasn’t. Real examples of how small missteps compound into crisis — and how to stop the bleeding early.

From late-night coding to open-source contributions — this session dives into why engineers keep creating even when no one’s asking. Real stories on how passion projects turn into breakthroughs and keep your engineering spark alive.

This talk breaks down how engineers are assessed at different levels, what “signal” really means in interviews, and why some candidates pass while others don’t. We’ll discuss how developers can position themselves to stand out, which skills matter most in today’s market, and how AI is reshaping roles and expectations. You’ll also get insights from Big Tech companies: the mindsets, behaviours, and practical examples that distinguish strong engineers from everyone else.

Successfully implementing AI is more than a tech challenge: it's a human one. This talk dissects the anatomy of failed AI adoptions to reveal a practical roadmap for success, focusing on the crucial interplay between technology, data literacy, company culture, and trust.

AI now powers healthcare, finance, and autonomous systems — but without explainability, it’s a black box we can’t fully trust. This talk unpacks modern XAI: from interpretable models to post-hoc tools like SHAP, LIME, and counterfactuals. Learn how transparency improves performance, fairness, and compliance — and why explainability is now essential, not optional.

The rapid adoption of AI tools is the biggest change in software engineering in the last several decades. But there are major differences in how the "AI-native" minority of teams use these tools, versus how the majority of the industry is experimenting with it. An overview on how AI labs and AI-first startups work with these tools, versus the rest of the industry; and what is working and what is not, in both places.

Ever catch yourself asking, “Should I swap GitHub for Outlook? VS Code for Google Sheets?” Same here. If that question keeps bothering you, pull up a chair.

In this talk we’ll cover:

• The switch: what really changes when you move from shipping code to leading people.
• Old friends, new roles: why yesterday’s teammate is now your direct report, and how not to make it awkward.
• The two-jobs paradox: when you know that your reports had another gig, but you can't say it.
• The hidden stuff: loneliness, calendar overload, tough feedback — all the things no one puts in the job ad.

Michael Lopp calls engineering management “an impossible job,” but it’s doable once you know the trade-offs. Come for honest stories, stay for practical tips, and leave knowing whether the manager path is right for you.

‍

‍

Is it time to refactor your career? I quit a comfortable job as an Engineering Director and spent over seven months traveling through nine countries. But don’t expect a slideshow bonanza — instead we’ll have a serious talk about decision-making. Travel is just the context.
Your career is the hardest system you need to design, and I’ll share my approach. Leaving is one big decision, but what comes after a long break? In 2025, the tech landscape changes daily — and so does hiring. How do you navigate complex dilemmas now, and how do you adapt to an AI-driven environment? We’ll look at what’s actually useful beyond writing resume-sending bots.

‍

In a world obsessed with tools, frameworks and automation, it’s easy to forget the human layer of technology. But the best products aren’t built by the biggest teams or the latest tools; they’re built by people. What if the real competitive advantage isn’t the next programming language or the latest technology, but the environment where engineers thrive? In this talk, Pipedrive’s CTO Agur Jõgi explores how happiness and collaboration fuel innovation and how Pipedrive’s “tiger teams” concept helps turn that philosophy into measurable impact.

You automate millions in savings yet stay invisible. You fix what breaks but lack the power to change what matters. This talk offers a practical framework to escape the cycle — how to push back strategically, communicate value through business impact, and turn DevOps from cost center to strategic force.

Large Language Models (LLMs) and AI-powered apps open new attack surfaces that traditional security testing misses. This session will demonstrate how AI Red Teaming identifies and mitigates those risks using real-world techniques and Microsoft’s PyRIT framework.

From a Warsaw-based startup to a global data cloud leader — this is the story of how Applica became part of Snowflake's journey to redefine enterprise AI. Join Michał to hear how a small team obsessed with AI grew into one of the engines behind Snowflake's Arctic models — and what it takes to scale from a research prototype to a platform serving thousands of data engineers. Whether you're a developer, data scientist, or founder - this is a raw, technical, and inspiring story about innovation, integration, and impact.

Engineering teams are entering a transition period: new tools, new responsibilities, and an entirely new skill set. In this panel, leaders from top tech companies will discuss how AI is reshaping engineering roles, which competencies matter most now, how to grow talent in this new environment, and what challenges teams must overcome to stay competitive.

As technology becomes capable of doing almost anything, the real challenge shifts — from how we build to what we build for. This talk explores the new architecture of product creation and how our processes must evolve to keep up.

How to design a vibe coding platform that combines the traditional editor approach with modern AI-driven capabilities, embedded directly into the Wix platform. Rethinking how next-generation developers can interact with code platforms and build software.

Being data-driven is no longer enough — the future is agent-driven. This talk shows how AI agents can move from analytics to autonomous execution, automating real workflows across departments. Learn from a live business case using Google Agent Space, OpenAI, and custom modules, and walk away with a roadmap to build your own digital workforce.

Emerging technologies — especially those driven by artificial intelligence and autonomy — are redefining the nature of military operations, strategic alliances, and the legal and moral frameworks that govern warfare. This talk explores the operational, ethical, and political implications of autonomous systems in modern conflict.

Innovation doesn’t fail for lack of ideas — it fails from poor management. This talk explores how to lead teams through uncertainty, balance autonomy and control, and build a culture where people experiment, fail safely, and turn ideas into lasting impact.

A focused conversation on practical ways to bring more women into technical roles — and why it matters for innovation and team performance. We’ll look at the systemic barriers that still exist, the programs and mentorship models that actually work, and how companies can build environments where women choose engineering paths and stay in them. Speakers will share real cases from their teams, showing how increased representation improves technical culture and business outcomes — and what the industry can do to scale this impact.

In iGaming, high load is the norm — rapid launches, constant pressure, and tech that evolves nonstop. Teams survive through tight coordination, backward-compatible releases, clean operations, and SLA treated as culture, not metrics. When every move affects the whole system, collaboration becomes the only way to deliver. With discipline and shared knowledge, pressure turns into a growth engine — ensuring every brand goes live stable, scalable, and future-proof.

‍

As AI rapidly accelerates execution and floods teams with endless ideas, the true edge for great PMs is judgment — knowing what to build, what to ignore, and where to place bold bets. This talk breaks down product intuition into a practical system: deep user exposure, pattern recognition, reflective learning, and fast feedback loops. You’ll gain actionable habits and frameworks to sharpen your instincts, make smarter decisions with incomplete data, and master product craft in an AI-driven world.

Transitioning from engineer to founder means trading code for strategy — without losing your technical edge. This talk shares lessons from building and scaling a startup grounded in engineering excellence and powered by AI. Learn how to turn technical skills into leadership strengths, balance coding with growth, and use AI to scale impact and business value.

AI tools can accelerate us — or make us lazy. This talk shares best practices I’ve tried for integrating AI into development workflows without losing your engineering edge, plus lessons from real cases where AI-generated code caused more problems than it solved.

The pendulum is swinging back: major tech companies are moving from microservices to monolithic architectures. This session explores why and how.
We'll examine the challenges driving this shift—from operational complexity to development velocity—walk through the journey from splitting a product into microservices to reassembling it as a monolith in a new form, using my experience and industry observations, and discuss the trade-offs and lessons learned. Learn when a monolith makes sense and how to structure it for scale.

AI models are under attack. From prompt injection exploits that bypass safety controls to data poisoning that corrupts training pipelines, adversaries are finding creative ways to compromise machine learning systems. This session examines the most critical threats facing AI deployments today: malicious prompt crafting, training data poisoning, sensitive data extraction from models, and supply chain vulnerabilities. Through real examples and live demonstrations, you’ll gain actionable insights into securing your AI infrastructure against these sophisticated attacks.

This talk introduces agency — the practice of acting intentionally, being a subject in the world. It's not about productivity hacks or motivation, it's about developing capacity to decide and move towards those decisions.

Agentic AI demos look magical, but real-world production demands *predictability*, cost control, and reliable results. I’ll share how we built an AI testing tool around a deterministic state machine (FSM/DAG) with structured specs, strict validation, and step-level observability — and why we chose this approach over free-form agent loops. I’ll also break down how patterns like Self-/Corrective-/GraphRAG and ReAct/Plan-and-Execute fit into this flow, and offer clear criteria for choosing the right one.

This talk explores how engineers can develop product thinking to drive business impact beyond technical excellence. Drawing from real experiences at Netflix and mentor.sh, I'll demonstrate practical frameworks for making engineering processes serve product goals rather than becoming bureaucratic overhead. Attendees will learn actionable strategies to balance technical rigor with product outcomes.

Product work changes radically as a company grows. What works at ten users often breaks at a million, and the balance between gut feel, data, and conviction shifts at every stage. This talk explores why some PMs thrive in one environment but struggle in another, and what truly defines success across different product phases. We’ll look at the lessons learned from building products for small, early audiences versus massive global ones, the hidden traps that appear along the way, and how to avoid them. Most importantly, we’ll discuss how to adapt your PM style to the company’s stage — and why this flexibility is now one of the most valuable skills in product management. Bonus - list of stupid ideas I pitched to my managers without thinking much.

‍

Building resilient services by assuming failures will happen and preparing for them. It walks through practical patterns like retries, circuit breakers, bulkheads, idempotency, and outbox, plus how to test and observe them. The goal is to design systems that degrade gracefully instead of breaking completely.

AI is everywhere today. Disruption is inevitable in, what seems like, all businesses. However AI will not implement itself in the companies - people will. How to make sure culture will not eat strategy for breakfast?

For decades, engineers built on certainty: databases return consistent results, APIs follow contracts, tests pass or fail cleanly. When an Airbus flight control computer receives a 15-degree bank angle command from the sidestick, it executes a precise sequence – calculating aileron deflection, adjusting elevator trim – and the aircraft banks exactly 15 degrees. Every. Single. Time. But LLMs have shattered this paradigm. The same prompt yields different outputs. Context windows affect behavior unpredictably. Your "API" might decide to change formats mid-conversation. As a lead engineer on Writer's Action Agent, I've been building production infrastructure where this non-determinism isn't a bug – it's the foundation.

This talk explores the profound mental shift from traditional fullstack engineering to AI platform engineering. How do you architect systems when you can't predict outputs? How do you build UI that renders agent responses restructuring themselves mid-stream? What happens when MCP servers give your LLM access to tools it can hallucinate. Through live agent development – MCP UI rendering, sandboxed execution that contains chaos, streaming architectures that adapt in real-time – we'll see the new patterns emerging: structured outputs as negotiation, validation layers as philosophy, prompt engineering as probabilistic API design. This isn't about learning a framework. It's about learning to think differently about what engineering is when your stack includes something that thinks back.

‍

This talk presents an architecture for running autonomous AI agents in production, combining Python for flexible agent logic with Rust as a secure, high‑performance "iron cage" runtime. You'll learn design patterns, monitoring approaches (including AI‑based quality gates), and a lead‑generation agent case showing how Rust limits risky behavior while preserving throughput.

Training your first ML model is easy — keeping it alive and improving is the real challenge. This talk walks through building an end-to-end MLOps pipeline: setting up CI/CD/CT, managing model versions, automating continuous training with Ray + GCP, and running production workflows efficiently and safely.

By 2030, cybersecurity for distributed systems will be less about restricting networks and software, and more about protecting data and keeping systems alive. As everything becomes connected, from drones and edge AI to microservices running across clouds, our software will need to make its own trust decisions, recover quickly, and exchange data securely even in messy, unpredictable environments.

In this session, we’ll talk about what’s actually working: Zero Trust, Data Centric Security, and Cross-Domain Security Solutions that make complex systems resilient by design. Expect current engineering practices, architecture patterns, and implementation lessons that are defining the next decade of cybersecurity.

In this talk, Artem will share practical cases of AI adoption in engineering teams: automating code reviews, generating production-ready code, creating prototypes, building internal AI tools, supporting engineers’ learning, and assisting in technical decision-making. These are not abstract concepts or internet examples, but real practices based on leading a 20-person frontend team across 7 products.

Shipping fast is table stakes now. The question is, how can your human and technology systems keep up? This talk looks at the engineering and organizational side of sustainable velocity: investing in outcome-driven platforms, using observability to stay grounded in data, reducing bottlenecks through clear organizational ownership, and empowering teams with AI tools to move quickly without compromising product integrity.

In the new era of agentic AI development, success depends less on experimenting with prompts and more on engineering structure. This talk explores how to bring software craftsmanship discipline into AI development by combining spec-driven development, custom skills (like those recently introduced in Claude Code), and proven technical design patterns. You’ll see how these approaches make your agentic AI development more predictable, robust, and effective without losing the creative power that makes it exciting.

In this interview, we uncover the remarkable story behind Photopea – a product built entirely by one developer, Ivan Kutskir. He created a world-class tool used by over 1 million people every day, without a team, without funding, and without corporate resources. We’ll dive into how Ivan built, scaled, and continues to grow Photopea on his own, and what every engineer can learn from his journey.

Most strategies die in slides after the offsite. At Preply, we learned their power lies in how they’re written, shared, and used every day. This talk shows how to make engineering strategy actionable and measurable — by involving engineers, aligning with business goals, and scaling it across diverse teams.

Many projects fail not because of lack of planning, but because key stakeholders are silent or overlooked, and conflicting priorities go unresolved. In this talk, I share real-life experiences of uncovering hidden stakeholders, aligning diverse goals, and turning silence into actionable collaboration. Attendees will gain practical strategies to identify invisible influence, engage overlooked voices, and drive projects to success.

Explore the technology and mission behind Waymo. This talk explores the journey of developing the Waymo Driver, a system of custom hardware and AI software designed to navigate the complexities of real-world driving. The talk will cover Waymo's stated goals of improving road safety and transportation accessibility for all.

This talk breaks down all key levels of database monitoring before diving into PostgreSQL specifics. Learn how to interpret vital metrics, compare popular tools, and avoid common pitfalls. We’ll also explore pgwatch — an open-source tool for quick, powerful monitoring — and show how to build anomaly detection and alerting on top of it.

At Ralabs, we built a framework to measure software quality that now tracks 50+ projects in parallel. Recently, we added AI agents to automate tasks like summarization and reporting, which reduced manual overhead and improved visibility. In this talk, I’ll share our approach, demo key parts, and highlight lessons learned on connecting quality with revenue and productivity.

Generating code, marketing content and Ghibli-style pictures with AI is all nice, but... there's also a dark side within this act of creation. We're still only humans, and "AI" is just a tool, therefore a human can ask AI to do something harmful, violent, or otherwise considered destructive, because... because why not? Jailbreaking is this act of "breaking the law" within the human-model interaction. And it's truly fascinating from both technical and ethical points of view.

Working across time zones requires more than good scheduling – it demands clear, empathetic, and culturally aware communication. This talk explores how asynchronous communication can improve collaboration, why feedback styles differ across cultures (as described in The Culture Map), and how to balance async work with meaningful face-to-face interaction. We’ll also cover practical tools like RACI matrices and calendar discipline – and why over-communicating is always better than leaving things unsaid.

The upcoming EU regulations introduce new requirements for companies using AI, particularly around explainability. In this talk, I will outline the specific provisions in the AI Act and the AI Liability Directive that will create mandatory obligations for AI explainability. I will explain their significance and the potential risks for companies that fail to meet these requirements. Finally, I will show how explainable AI techniques can be used to address these regulatory obligations.

Agents are increasingly everywhere, but building one that actually works in production is a very different challenge than building an initial demo. What does it mean to design an agent that can reason, use tools, stay on track, and run for hours without becoming unpredictable or fragile?

In this talk, we’ll break down what agents are, how they’re put together, and the practical decisions that matter most when building them. From prompts and models to harnesses, tools, evals, and guardrails. You’ll walk away with concrete best practices and mental models you can apply immediately to build agents that are reliable, observable, and ready for real-world use.

This talk shows why Vibe Coding unlocks real value for Product Owners, yet rarely brings the same impact to engineers. It offers a clear, grounded view of how AI fits into product development: when it speeds things up, when it creates friction, and where its limits are. The focus is on a sane, practical way to use AI that aligns with product intent and leads to faster decisions and shorter time-to-market. No hype, just observations and lessons from real projects. This session is for Product Owners and engineers who want to improve the entire development cycle, from the initial idea to delivering value to the end user.

A straightforward conversation between founders about what really pushes specialists to start their own companies — the first steps, the challenges, and the decisions that matter most. No host, no fluff — just practical insights from people who’ve done it.

Many teams ship hand-written or AI-generated API clients that duplicate types, drift from API contracts, and slow development. This talk shows how to generate SDKs directly from OpenAPI/GraphQL/TypeSpec contracts or even backend source code to improve developer experience, speed, and reliability. A live demo walks the workflow end to end and include a real case how one team cut breaking-change fixes from days to minutes. After this talk, you will be able to replace manual client layers with repeatable, consistent contract-based SDKs and keep them in sync.

In this session, Vitaliy Goncharuk will outline the key technological shifts—like AI and robotics—and the evolving geopolitical dynamics between the EU and U.S. that are reshaping the defense and dual-use sectors. These trends are fueling strong demand for engineering talent. Against this backdrop, he’ll break down practical career strategies, hidden risks, and overlooked opportunities—especially for developers and engineers from Eastern and Central Europe looking to enter or advance in the field.

‍