Hello, cloudspaces.gr

After years of working in data and infrastructure, I decided it was time to write publicly.

Not because I have all the answers. Because I spend a lot of time working through hard problems, and I’ve learned more from other engineers’ honest accounts of what went wrong than from any polished documentation.

What I do

I’m a platform engineer focused on data infrastructure and, increasingly, ML platforms. That means:

• Operating Apache Kafka clusters that process millions of events per day
• Managing Hadoop/HDFS at scale on Cloudera CDP
• Running Kubernetes clusters that serve as the compute layer for data workloads
• Building the CI/CD tooling that lets data engineers ship pipelines without filing ops tickets
• Learning MLflow, Kubeflow, and SageMaker — because the data platform naturally evolves toward serving ML

My background is 10+ years in Linux and networking. That foundation shapes everything: I care deeply about observability, I’m suspicious of complexity I don’t understand, and I believe a system that only works when nothing goes wrong isn’t really working.

What I’ll write about

Mostly practical things:

• Hard-won operational knowledge about the Apache ecosystem
• Architecture decisions with real trade-offs, not just “it depends”
• Notes from learning ML platform engineering as an infrastructure engineer
• Occasional takes on the state of data tooling

I’ll try to be honest about what I don’t know. There’s too much confident content written by people with 6 months of experience. I’d rather be specific and occasionally wrong than vague and always safe.

The name

cloudspaces.gr — the intersection of cloud infrastructure and the open space of ideas. Also it was available and I’m Greek.

If something I write is useful, wrong, or worth arguing about — send me an email.