Milen Dyankov
A Future-Proof Architect in a Shiny AI World
Łódź, Poland
Actions
Milen is on a mission to help software developers worldwide design clean, modular, and future-proof systems. With over 25 years of experience building distributed architectures, consulting for global tech companies, and leading engineering teams, he brings a "software craftsmanship" lens to modern development. While his roots are in enterprise Java and complex systems, Milen is currently focused on navigating companies and developers through the "AI ocean" without drowning. He is dedicated to ensuring they continue to build well-architected solutions rather than falling for the newest, shiniest products.
Links
Area of Expertise
Topics
Location transparency
If it is the first time you've heard the term, then this talk is for you! This talk is definitely for you if you are concerned about Coherence, Coupling, and Connascence metrics while building applications. Perhaps the amount of things that can break prevents you from splitting a monolith into distributed micro-services? Applying location transparency may be both your kick-starter and your safety net.
This talk will examine a traditional application with its dependencies between methods, classes, packages, and modules. We'll talk about how those impact application's ability to grow and be decomposed into independently deployable units. Then we'll see how and where to apply location transparency. Finally, we'll rise above the code to explore how Domain-Driven Design, CQRS, and Event Sourcing architectural concepts tackle complexity and why location transparency is key while doing so.
Give your application memories
Most applications only know the present. Their current state is calculated and updated somewhere on every change, effectively losing any previous states and information that only existed in the past. Can that be a limitation? Missed opportunity? What if your application could remember its entire past? Would that be any useful, or is it a worthless effort?
This talk will provoke Java developers to start asking such questions. It will provide useful mental models to better understand Event Sourcing, DDD, and CQRS concepts. It will also demonstrate what it takes to build such applications in Java.
Demystifying "event” related software concepts and methodologies
No matter exactly what type of software you are building, at least a few terms starting with the word "event" must have tried to invade your professional space. "Event-driven" has already evolved to the level of a magical buzzword in IT marketing and sales activities. But the tech terms list is much longer. It includes "event streaming", "event storming", "event sourcing", "event store", "event modeling", "event notification", "event-carried state transfer", ... Getting dizzy already? Don't worry. Many developers feel this way. This talk is an “event-dizziness” cure. It'll walk you through the entire "event land" and explain what is what. It’ll discuss the origins, intention, and applicability of each term. So the next time you face the "event-driven" label, you can ask the right questions and make educated choices.
AI for Java Developers - From Buzzword to Code
AI is undeniably a buzzword, but it is here to stay. What does this mean for Java developers? Can we design and build applications that harness the power of AI? Most importantly, where do we start?
This session focuses on inference—using pre-trained models directly from your code. Instead of relying on proprietary APIs, it'll demystify the "magic" by breaking the AI buzzword down into familiar concepts and components. This will enable you to use AI models as naturally as you would work with databases or workflows. Through live demos, you'll learn to leverage basic embedding models, explore vector databases, and use tools like LangChain4J. By the end, you'll have a practical roadmap to kickstart your journey into Java-powered AI.
Adding "Simple" AI Features to a Movie Streaming Platform
This workshop-converted-to-talk shows extending a sample movie streaming platform with semantic search, recommendations, and similarity explanation. It demonstrates vector search and LLM integration. While the code you'll see doesn't come from Netflix, it's an example of how embracing relatively simple AI tools yields impressive results.
You'll learn how to chunk text, why combine dense and sparse embeddings, how vector centroids capture user preferences, and what it takes to merge multiple search strategies. You'll also witness the implementation of a RAG pipeline that finds similar movies and explains those similarities.
The presentation will not blow your mind with the next groundbreaking AI achievement, but will show you code you can understand and write yourself. The focus is on making AI features work reliably and delivering results rather than showing off.
DevBcn 2025 Sessionize Event
Milen Dyankov
A Future-Proof Architect in a Shiny AI World
Łódź, Poland
Links
Actions
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top