Thank you for pointing that out and for your consideration.
In the Japanese version, sentences and dialogue are often written without explicit subjects, so when they are translated directly with DeepL, the subject is frequently and automatically rendered as “he.” To avoid this, I tried to add subjects before translating, but I believe there are still many parts I didn’t manage to catch.
However, fixing all of those instances would require a significant amount of time and effort, and it would slow down my current work, so for now I’ve decided to leave them as they are.
If you notice similar issues in my next project, please let me know. I’ll try to correct as many of them as possible.