/images/avatar.png

Tech Lead and AWS Developer

Three Pipelines, One Platform — My CI/CD Architecture on AWS

In the first post I described how aws-sensei.cloud is set up — Hugo, S3, CloudFront, a CodeBuild pipeline. That was a good start. But one pipeline for everything doesn’t scale. The problem showed up in practice faster than expected: with a single pipeline, I kept hitting the CodeBuild free tier limit just by writing blog posts — even though I had only changed Markdown. Every commit triggered the full pipeline: Hugo build, infrastructure deploy, everything.

From Vercel to AWS — Migrating My Blog to the Cloud

I have been working with AWS for over six years — in projects, in architecture, day to day. At some point it became clear: if I take my AWS knowledge seriously and want to show it to the world, my own infrastructure should reflect that. That is how AWS Sensei came about — as a platform to share knowledge, try things out, and make AWS skills visible. And if you are advertising your AWS skills, your own blog should run on AWS too.

Fake Methods and Dependencies without Dependency Injection

Mocking via virtual method Fake Dependencies with an Interface and Dependency Injection is easy and the usual approach. But how can you fake without Dependency Injection or an Interface. I will show you two situations that came up in production code. Factory Method The first situation is a small AWS Lambda function with an http call. We want to mock the .Net HttpClient and write a UnitTest for our Request method.