This is a good example of how unscientific thinking works.

Ketosis works well for epilepsy,especially treatment resistant or treatment incompatible such as in kids. It also seems to help with depression and anxiety, along with showing promising though not yet conclusive results for supporting ADHD and some difficulties with ASD. In those cases we see lots of overlap with schizophrenia in terms of family history and symptom presentation, so they seem related. A reasonable enough conclusion is that if something helps with one of those it may help with the others.

A good example is sodium valproate. It is an anti epileptic medication and helps prevent seizures. It also helps with bipolar disorder. Is this because they have a shared underlying mechanism? Is this because the drug has more than one effect and both disorders are benefited by valproate? Is there a third thing that valproate impacts which impacts both disorders? We don't know.

Ketosis helps with a bunch of things but has not been shown to be useful for schizophrenia. It is possible it is the treatment of the century, a wonderful intervention which will make problematic schizophrenia symptoms a thing of the past. Is that likely? I don't know and neither does he. We simply don't have evidence for that yet. We don't know if it is useful for schizophrenia or how useful or in what context. A scientific thinker would say that. A lawyer who wins not be making a scientific argument but by convincing people doesn't care about evidence, he cares about making a reasonable enough argument and sounding authoritative while doing it.

It is not science based. He is a crank. If he tells you the sky is blue check for yourself, but don't Google "is the sky blue", search Wikipedia for "sky" and read the whole article.