A Realistic Day in my life as a Software Engineer in Germany

🎵 I started the morning with a run and now I'm working from WeWork in Munich. Today I'm taking you through a normal work day as a software engineer in Munich. Most of my work today is going to be about a major rebrand that affects many parts of the product that I'm working on and also I have a couple of meetings in between. This video I'm going to show you what I'm working on, how my setup looks like or dev setup looks like and how a day in the life looks like as a software engineer today. So, stick around till the end to see what happens. The design system that we currently use has undergone an update. That means I need to bring the changes into many domains of the product that I work on. Some screens usually use the components from the design system while others use custom made components. I also need to make sure that if I could reuse some of the new design system components rather than having a custom one. In the end I verify how everything looks after the migration and provide results of the before and after screenshot. It kind of sounds like a visual task and yes it is a visual task but a lot of things are very hard to understand from the get go. And since I'm starting with this product for the very first time, a lot of unknowns and a lot of things I had to understand before I could make any change. Even though AI is helping me a lot, there are a couple of things as someone who is new to the team couldn't really understand how it works. One tricky part of this rebrand is that my team doesn't own the design system. Another team updates it and our product uses those components. So change from their side can affect something I've already migrated. Before anything goes out to the customers, I need to make sure the existing screen still look right and work the same way. I just want to show you guys how my current agent setup looks like. I'm moving through the product one domain at a time and Cloud Code is my main coding agent. I keep one terminal for rebrand work in that domain. A second one handles code coverage changes. If I'm still learning that part of the product, I'm open to a third one for documentation. Most of our documentation is in Notion. Cloud is connected to Slack, Linear, GitLab and Notion through the company's MCPs so I can pull the context into one single place. I haven't been typing for quite a while now. I've been using this tool called WhisperFlow for most of the instructions because talking is much faster for me than typing all of that. The downside is that all of this can get distracting very quickly. While one agent is working, I move to the review. Then the first one comes back with a question and I have to remember what I was doing there. It feels a bit like having too many Chrome tabs open except every tab might be changing code. So I normally stop at two or three contexts. The actual work, review and the documentation when I need it. So there's one thing that AI still misses. The agents can read code and search the documentation but the one thing they're not so good at is identifying the typography differences or the color changes or anything to do with the padding or how the elements flow in the webpage. Even if you provide them the screenshots, they don't have much context on how the actual output should look like. I still need to manually test the page and speak to the people working on it to understand how the final outcome should look like. I've been using these agent setups for my own puzzle projects. I've also been testing some of the AI workflows but parts of the videos have been edited by AI. It arranges the clips, removes obvious gaps and filler words and puts a rough version on the final cut timeline and probably it saves me around one to two hours on every video. Let me know if you guys want to see how I edit those videos or how I use these agents to help me edit those videos. I can cover that in the next video.