Remote machine learning on Windows with Docker and WSL2 from anywhere

Alex Kretzschmar from Tailscale provides a guide on utilizing a personal Windows 11 gaming PC for running self-hosted photo and video management services, specifically using Immich, by leveraging Docker and WSL2 with Nvidia hardware acceleration. The process allows users to perform tasks such as facial recognition and smart search object detection on their own hardware, keeping data private and under their control. The video elaborates on connecting machine learning components to the system and extends the concept to include any GPU within a tailnet, even those of remote friends. This approach aligns with Tailscale's mission to help users create small, trusted networks, and encourages engagement and feedback through various social platforms.