Talks

How to implement a native Rust engine in a Java application via Project Panama (FFI). Practice and pitfalls.

Let's analyze and compare the current garbage collection algorithms in Java and Go.

I'll show you different Outbox implementations.

I'll show you Nullaway, an extremely smart static nullability analyzer.

With the help of Claude Code, we will write a personal database that will not be inferior in performance to leading solutions.

Tired of reviewing the code? Connect the AI! But don't be surprised if he starts spamming. 

From a failed pilot to a production agent: how we created an automatic code review system for Spring AI. I'll show you why the most sophisticated prompt engineering is not an option, how architecture and limitations make AI "smarter" and how we taught the bot to leave only really useful comments.

Don’t wait for a 3 AM production alert to find your race conditions. Master JCStress, Lincheck, and Fray before the users do.

I'll tell you how we eliminated a serious drawback of serialization with the One Nio library: ignoring custom methods designed for standard Java serialization.

I will also tell you about the elimination of one of the limitations of the One Nio serialization algorithm, which made it possible to do this even faster.

We will tell you how we provided the agent with the execution time data and what came of it.

A talk on the evolution of the S3-compatible storage metastore for billions of objects, architectural patterns and errors along the way. How background operations and statistics collection are performed efficiently.

A talk on optimizing write performance in Apache Cassandra.

Let's look at why the Spring MVC service can freeze when the CPU is low: the queue accumulates not in the processor, but in Tomcat/Hikari/HTTP pools and waits for external dependencies. You'll get a minimal set of metrics (RED+USE), a short diagnostic playlist, and practical fixes: timeouts, bulkhead/limits, pool matching, and saturation alerts.

I'll tell you about the migration of the service that we did.

Advice that various AI assistants give on improving the code.

We want to talk about each stage of increasing the load and bring the service to a situation in which there are no standard solutions anymore.

In the talk, let's recall what value classes are in general and why they are needed in Java, and what tasks they will be useful for.