Programming
Faster, Leaner, Predictable: AI Microservices on Quarkus vs Spring
Alexandru Diniș - Unit Head Cluster Engineering @ BMW TechWorks Romania
Studio room
12th November, 11:30-12:00
AI features live or die by startup time, memory footprint, and tail latency. In this talk, I implement the same text-classification service (DJL + Hugging Face) in both Quarkus and Spring Boot, then compare developer experience and production metrics: JVM vs native startup, idle/loaded memory, p95/p99 latency, and container image size. I’ll also show how the BCE architecture keeps the code portable and testable. You’ll leave with clear trade-offs and a checklist for choosing the right Java framework for AI inference in the cloud.
Alexandru Diniș
BMW TechWorks Romania
I started my tech journey 12 years ago as a Java developer and learned the stack the hard way: keeping up with Java versions, adapting to new frameworks, and moving from solid monoliths to modern, distributed systems.
Most of my work has been in the automotive field, where reliability matters and the smallest details add up.
Over time I moved into solution architecture and delivery management, making sure ideas turn into real, shipped software.
Today I lead a group of 50+ engineers, focused on people development, writing clean and efficient code, collaborating with stakeholders, and delivering results that make a difference for the business.
I’m hands-on, passionate about Java, Cloud and AI solutions and about finding new ways to integrate them into the development lifecycle, from design to delivery. I’m also deeply interested in DevOps practices and security. Beyond that, I enjoy exploring, writing about what I learn, and sharing practical notes and lessons from the field.
