I Gave a LLM Keyboard ⌨️ Pray for Me 🙏

We've spent years teaching LLMs to talk. Now let's give them hands

What happens when your friendly little text generator can use tools, run code, poke at a real environment, and actually do things?

In this quick whirlwind, we'll watch an LLM go from:

"Here's some code."
to
"I ran it."
to
"…uh, why did it do that?"

We'll play with tool calling, structured outputs, and Docker sandboxes to see how you can give an LLM a tiny disposable computer of its own — enough freedom to be useful, enough guardrails to (hopefully) keep things interesting.

No 47-slide AI manifesto. No endless architecture diagrams. Just a fun little peek at what happens when you stop chatting with your LLM and **put it to work.

Bring your curiosity.

I'll bring the keyboard.

Understanding

• Sandboxing

• Guardrails

• HITL (Human in the Loop)

• Agentic Architecture