The Drop: Making LLMs More Predictable with Boundary’s BAML

Working with large language models can feel fragile.

You write a prompt. It works.
You change one thing. It breaks.
Your code expects structure. The output does not cooperate.

This first Drop is about how Boundary’s BAML reduces that fragility by adding structure and clearer expectations to AI workflows.

We are not here to explain how LLMs work.
We are here to show how to work with them more safely.

What we will focus on:

• How BAML helps define what an LLM should return
• How structured outputs make AI responses easier to use in code
• How BAML fits into simple Python workflows
• Common patterns that make AI code easier to maintain
• Practical ideas worth stealing for your own projects

No advanced AI knowledge required. If you have used prompts, APIs, or AI tools, you will immediately recognize the problems.
.
Who should attend:

• Beginners experimenting with AI
• Data analysts and data scientists
• Software engineers curious about AI tooling
• Anyone tired of unpredictable AI responses breaking their code

Beginners and veterans welcome.
No gatekeeping. No dense decks.

Agenda:

• 5:30 to 6:00 Announcements and introductions
• 6:00 to 6:45 BAML and Python walkthrough and discussion
• 6:45 to 7:00 Q and A, and what is next

The goal is simple. Fewer surprises. More reliable workflows.
Leave with a more straightforward approach and at least one idea worth stealing.

This is the first Drop. Come help set the standard.