Andrew Burke focused on improving developer experience and documentation quality across PyTorch-related repositories. In pytorch/audio and pytorch/vision, he standardized nightly installation instructions by replacing outdated Conda references with Pip-based guidance, ensuring alignment with the official PyTorch site. This work, implemented in Python and Markdown, reduced setup friction and improved consistency for contributors. Later, in meta-pytorch/forge, Andrew overhauled the TorchForge onboarding documentation, clarifying system requirements, installation, and verification steps while removing obsolete content. His technical writing and build systems expertise enabled faster onboarding and more predictable build outcomes, demonstrating a thoughtful approach to maintainability and user-focused engineering solutions.