Coding LLM APIs — From Cloud to Local

In this workshop, we’ll write real code to interact with both proprietary LLMs (like OpenAI and Anthropic) and open-source local models (like those run via Ollama).

This is your chance to move from conceptual understanding to building working prototypes using Python and REST APIs.

1. Coding a Proprietary LLM API (OpenAI / Anthropic)

• Set up API keys securely
• Write Python code to send a prompt and receive a response
• Handle parameters (temperature, max tokens, etc.)

2. Coding a Local LLM API with Ollama

• Start an Ollama server and test locally
• Use Python (requests or similar) to send prompts
• Compare performance and output with proprietary APIs

3. Bonus: Wrapping It in a Simple App

• Create a basic CLI or web interface (e.g., Streamlit or Flask)
• Think about extending the code for your own projects