From Python scripts to LLM applications
What changes when a print/input program becomes an API that calls a model — tokens, retrieval, and honesty.
Objectives
- Separate classroom Python from production CPython
- Explain tokens, cost and context windows
- Choose RAG when facts change often
Skills: Python · APIs · RAG
The honesty gap
Sharda classroom Python runs print, input, int, if, range and lists. Production AI apps use packages, HTTP, secrets and evaluation. Finish the Python course for the first habits, then treat LLM work as a new engineering layer — not a certificate that you are an ML researcher.
Ground answers
An LLM predicts tokens. RAG retrieves current documents at ask-time so you can cite them or refuse. Fine-tuning changes weights and is the wrong first tool for a handbook that updates monthly. Log retrieval, measure faithfulness, and never paste secrets into a prompt.