recommenders-team/recommenders ? reverse-engineered prompt

Build me a Python project for recommendation systems that feels like a practical learning and experimentation toolkit.

I want it to help someone go from raw user and item data to a working recommender, with clear examples for preparing data, training a few different recommendation models, evaluating them, and tuning them. It should include notebooks or examples that are easy to run for common cases like collaborative filtering, sequential recommendations, and content based recommendations, plus a simple quick start that shows the whole flow on a sample dataset.

Please make it useful for both trying things out locally and for moving toward production, with helpers for loading data, splitting train and test sets, and checking model results. If there are optional extras for GPU or Spark, keep those in mind too. Keep the code clean, well documented, and easy to extend with new algorithms later. If you need to look up current library docs or best practices while building it, go ahead and do that.