LLM Zoomcamp 1.2 — Environment

Why It Matters

A consistent, remote environment and proper dependency/secret management reduce onboarding friction, avoid reproducibility problems, and help control API costs when building and testing LLM-based apps. Proper setup streamlines collaboration and troubleshooting across learners or teams.

Summary

The instructor walks through preparing a reproducible Python environment for the LLM Zoomcamp using GitHub Codespaces and VS Code, creating a new repo and initializing the project with UV to produce a pyproject.toml. They install key dependencies (requests, a vector-search library, and the OpenAI client), create a Jupyter notebook, and configure .gitignore plus a .env for storing an OpenAI API key. The instructor recommends Codespaces for consistency (Colab is possible but limited), demonstrates selecting the kernel and running a test cell, and advises creating a dedicated OpenAI project to track usage.

Original Description

Setting up the project with Python and uv, installing dependencies, and configuring your OpenAI API key (plus OpenAI-compatible alternatives like Groq).

Lesson notes: https://github.com/DataTalksClub/llm-zoomcamp/blob/main/01-agentic-rag/lessons/02-environment.md

LLM Zoomcamp is a free course on building real-world LLM applications: https://github.com/DataTalksClub/llm-zoomcamp

Module 1 (Agentic RAG), Part 1 — lesson 2 of 10.

Chapters:

0:00 Preparing the environment

0:10 Creating the project in Codespaces

3:26 Installing uv

3:59 Adding dependencies

6:29 Creating the Jupyter notebook

8:09 Setting up .gitignore and .env

8:43 Getting an OpenAI API key

11:02 Loading .env and the client