I’ve been thinking seriously about my next career move, and I’m honestly a bit confused about where to put my effort.
My background is mainly Data Engineering + DevOps. I’ve worked with AWS and Azure, built data pipelines, worked with CI/CD, Jenkins, GitHub Actions, automation and cloud infrastructure. Kubernetes is something I haven’t used extensively in production yet, but I’ve spent a lot of time understanding the architecture, workloads, networking, deployments and how the whole ecosystem fits together.
I currently work around large enterprise systems in the automotive space, so my day-to-day thinking is very much about reliability, automation, deployments, data and infrastructure.
Now I’m looking at AI Platform Engineering / MLOps / ML Infrastructure.
And this is where I’m stuck.
I’m not particularly interested in becoming the person who trains the next foundation model from scratch.
What actually interests me is everything around it.
How do we build the infrastructure that ML engineers and data scientists depend on?
How do we deploy models reliably?
How do we manage GPUs?
What does CI/CD look like for ML?
How do you monitor models in production?
How do you handle model and data versioning?
How do you build an internal platform where someone can take a model from development to production without having to fight infrastructure every step of the way?
The problem is that once I start researching this space, I end up with another giant list of tools.
Kubernetes. MLflow. Kubeflow. Ray. Terraform. Airflow. Databricks. SageMaker. Azure ML. Vector databases. LLMOps. RAG. GPU infrastructure...
And that's exactly what I don't want to do.
I don't want to spend six months collecting tool knowledge and then realize I still don't understand how a real AI platform is designed.
So I'm trying to figure out what the smartest transition looks like from Data + DevOps → AI Platform Engineering.
If you're already working in this space, I'd genuinely like your opinion:
What would you learn first if you were starting from my background?
And more importantly, what would you deliberately NOT learn yet?
If you're making a similar transition, I'm also open to studying together and building some real projects around this instead of just following another 50-topic roadmap.
Would be great to hear from people who've actually done this transition.
Top comments (0)