Technical Deep Dive: How a 30-Year-Old Syntax Became the Most Powerful Tool in Your AI Stack

​​A valid SXSW badge is required for entry.

​​Large language models were trained on vast amounts of structured web text, much of it written in lightweight markup formats like markdown. In this session, we examine why markdown-native design aligns so closely with how transformer models process and generate text — and why that matters for teams building real-world AI systems. Through live demonstrations and practical examples, we’ll compare traditional system prompts with fully structured markdown approaches, explore how markdown improves clarity and controllability, and share implementation patterns for building, versioning, and scaling AI systems in production.

Andrew Demers — Senior Technical Product Manager, webAI

​Andrew leads product strategy and execution for webAI's decentralized AI platform, with a focus on AI infrastructure, on-device inference, and distributed network architecture.

​​Official SXSW House Event

​​A valid SXSW badge is required for entry.
Each session is capped at 60 guests to keep the experience focused and high-level.
Approved RSVP + QR code required at the door.
Doors close once capacity is reached — please arrive early.