Worked on the lightly-ai/lightly-train repository to enhance developer onboarding and clarify project capabilities through comprehensive documentation updates. Focused on technical writing using Markdown, the work included adding detailed benchmarks for object detection and semantic segmentation models, enabling objective performance comparisons for users. Installation instructions and tutorials were revised to streamline setup and reduce onboarding time, directly supporting faster adoption and clearer user expectations. No bug fixes were addressed during this period, as the primary emphasis was on documentation quality. These improvements support data-driven model evaluation and align engineering documentation with business needs, ultimately reducing support overhead and improving product usability.