I have seen a lot of papers and such on ways to optimize serving (usually language) AI models. Is this a space that is approachable to learn for people who are not interested in becoming experts on the actual foundation model architectures? If so, is there much beyond papers to learn from at this point - and do you see this becoming a subfield (similar to e.g graph database systems) that people will be a source of domain experts in the future?
Ask HN: Learning Model Inference Infra | Better HN
