Dynatrace Innovation & Technology Culture

What types of products or services does your engineering team work on/create? What problem are you solving for customers?

Our app team creates an internal app for Dynatrace internal developers who are our “customers.” When other teams want to release an app on the Dynatrace platform, they will soon all need to release it via our Dynatrace Console. Think of it as a content management system for app releases, where people can put their marketing information about the app, add screenshots, add links for more details or add related apps on the platform. Developers can also manage their app releases on the console, being able to unpublish faulty releases or manage the changelog of a specific release. Our Dynatrace Console will soon be integral to the latest Dynatrace Apps platform release process. We’re currently in the middle of migrating existing apps to the Dynatrace Console. We want to make it as easy as possible to allow a smooth release process for our internal app teams at Dynatrace.

Tell us about a recent project where your team used AI as a tool. What was it meant to accomplish? How did you use AI to assist?

We were in several pilot phases where we could try out new AI coding tools. One of them was GitHub Copilot. When using these tools, I noticed that there is not a lot of training data included in the models about our internal design system called “Strato Design System,” so I decided to write a small Model Context Provider, which provides documentation data from the Strato docs to the LLM. Now, with the help of this tool, we can significantly improve the quality of our LLM-assisted coding sessions because the AI agents can retrieve the correct usage information about our internal design system components. With the help of this context, the LLMs can produce a vastly better coding output. 

Just recently, I had to port some of our data tables to a new version of the Strato data table, which would have been a cumbersome task, but with the help of coding agents and the Strato docs MCP, the agent did most of the work. I could then go over the changes, refine them and get the migration through instantly. This would’ve taken me a lot of time otherwise, and the agent could do the heavy lifting for me here, which was a big win for us.

What would that project have looked like if you didn’t have AI as a tool to use? How has AI changed the way you work, in general?

AI-assisted coding can be very helpful, but I think the key is adopting it with intention. To me, it is important to never give up on the agency over your code. Never let the AI do things that you don’t understand. Never senselessly let the AI generate loads of code. Use this new tool like a new one in your toolbox, and learn how to use it properly. My approach to AI-assisted coding is to have a clear idea of how to solve coding problems, let it do the cumbersome work, but know exactly what I, as a developer, am doing and what the AI agent is doing. Give the agent a clear context, give it a clear purpose and approach it with a solution in mind. 

Having all of these best practices in mind, I have to say that the general use of coding agents has improved the speed and output of code without degrading the code quality or the general reliability of our software. I can definitely say that AI has changed how I work, though not in a way that I let it do my work, but more like I have found a co-worker that never gets tired of my questions and is always up to spar with me and talk about technical problems and their solutions.