Hugging Face releases playbook

Hugging Face consolidated a set of 12 free, self‑paced Learn courses that span LLMs, robotics (LeRobot), Model Context Protocol (MCP), agents, diffusion models, 3D ML, audio, computer vision, deep RL, ML for games, an open‑source AI cookbook, and a compact "smol" post‑training course. (huggingface.co) The MCP course was produced in partnership with Anthropic and includes a hands‑on MCP server, SDKs for Python and JavaScript, Gradio integration, and guided units that teach building end‑to‑end MCP applications. (huggingface.co) LeRobot—the open‑source robotics library used across Hugging Face's robotics course—provides PyTorch implementations, a LeRobotDataset format for sharing recorded demonstrations, pretrained models, and simulated environments designed for imitation learning and reinforcement learning workflows. (huggingface.co) Several courses embed quizzes, step‑by‑step units, and "Get Your Certificate" completion flows, while the MCP and robotics tracks surface runnable examples and hosted servers so learners can submit code and experiment with live tool integrations. (huggingface.co) Hugging Face has also surfaced partner learning tracks and project pipelines (for example a Hugging Face Fundamentals skill track on DataCamp and third‑party specializations using HF tooling), which are explicitly framed around building project artifacts and real‑world demos suitable for resumes or internship portfolios. (huggingface.co) Source code, notebooks, and reproducible course material are published in public repos and Spaces (robotics‑course on GitHub, MCP course README and a Smol Training Playbook Space), enabling direct forking, deployment, and inclusion of running demos as portfolio evidence. (github.com)