The author suggests that those working on this aspect of LLMs are more akin to API/backend engineers, as they need to understand the specific AI/LLM data model and its use cases, but they are not the ones who originally engineered the model. That task falls to R&D engineers, who the author refers to as AI engineers. These AI engineers build the data models that others then use.
Key takeaways:
- A web app engineer's perspective on LLMs involves taking customer text input, calling APIs, and delivering results over HTTP.
- LLMs ingest and organize text data, and provide APIs to output text based on certain inputs, similar to working on Elastic or other search engine technology.
- The APIs being called likely maintain state by keeping track of inputs, context, and outputs.
- The role of an API/backend engineer in this context involves understanding the AI/LLM data model and its use cases, but not necessarily engineering the data model themselves. This is typically the job of R&D engineers, who are referred to as AI engineers.