Ask HN: Sources / guides for custom model training and evaluation stacks

Are there any definitive sources / guides for the stacks companies are using as they use their own datasets to build & train custom AI models? I'm thinking the high-level workflow for data discovery through to labelling / creating ground truth to model evaluation.

I know it's moving quickly and I've seen various tools and approaches from different enterprises. I'm wondering if there's any standard starting to emerge or a place where teams are sharing tools / approaches so I can learn and participate.