Gen-AI Developer Classroom notes 18/Oct/2025

Programmatic LLM Interaction setup

• Lets create a new folder and cd into it

mkdir basic
cd basic

• Now execute the following command

uv init .

• Lets execute the following command to get the virtual environment

uv sync
# mac or linux
source .venv/bin/activate
# windows
.venv/Scripts/activate

• Now launch visual studio code code .

• Ensure the following extensions are installed

  • python (from microsoft)
  • jupyter (from microsoft)
    

• Lets add the following packages

uv add langgraph langchain

• Create a file llm_experiment.ipynb
• Now Execute Ctrl + Shift + P in windows/linux and Cmd+Shift+P
  • >Python: Select Interpreter
    • Select the Virtual environment in local directory .venv/...
• Now in jupyter notebook select kernel with Same python environemnt as above and create a cell write a simple python code print("hello") and Exectute.

• Install Kernel will be prompted once for every new environment

• Lets set up for GCP Vertex models

uv add "langchain[google-vertexai]"

• Refer Here and select google vertex

LLM Interaction Credentials

• Generally LLMS will have either
  • API Keys
  • Configuration Steps to create a file which will be used by default

Google Models

• We have 3 options
  • Google Gemini
  • Google Cloud Refer Here
  • Open source Google Models running on your infra (only Gemma Models)

OpenAI

• We have 3 options
  • OpenAI
  • Azure OpenAI
  • Open source OpenAI Models running on your infra (gpt-oss models)

Anthropic Claude Models

• This has many options
  • Anthropic
  • Azure
  • Google Cloud
  • AWS Cloud

Setting up Authentication with LLMS

Google Cloud

• Steps:
  • Install google cloud cli
  • gcloud init
  • gcloud auth application-default login