Review architecture, models, and design of semantic search system

[mlissner]

We have done a lot of work towards our semantic search engine, including selecting models, designing the architecture, and building a microservice (named inception), to do the embedding process.

Rachel is our new AI engineer, so she needs to get up to speed on all this. This ticket is for her to do that, ask questions about what she sees, and identify anything she'd suggest that we might want to do differently. (Hopefully that's not much!).