ProjectBuild an Agentic Content Studio: the explanation · Module 3: Choosing the model and the infrastructure · part 5 of 22
Part 5 · Module 3: Choosing the model and the infrastructure
Module 3: Choosing the model and the infrastructure
22 min read·
About 45 minutes · Needs: Module 2 (the solution) · Code: research/spike_llm.py, backend/app/ai/llm.py, backend/app/core/config.py, docker-compose.yml
Where this fits
The model and web search (amber: outside our code, and not always available) and the infrastructure around them: nginx, PostgreSQL and the containers.
Every agent in the studio calls one model through one small client. This module explains which model was chosen, where it runs, how fast it is, what it cannot do, and how the rest of the infrastructure (nginx, PostgreSQL, the containers) is set up around it. These choices decide how long a user waits and what the code must do to cope.
The rest of this project is yours to keep
This project is bought on its own, once, and stays readable afterwards, including the parts added to it later.