Building Robust AI Systems with Battle-tested Frameworks - Mini Summit #10

Join this MLOps Community Mini Summit featuring two insightful talks on production AI systems. Learn from Boundary ML's CEO Vaibhav Gupta about "Breaking the Demo Barrier and Getting Agents Shipped," where he shares proven strategies for keeping Large Language Models robust in production environments, including structured prompting, dynamic routing with fallback handlers, and data-driven guardrails. Discover why naïve JSON usage can reduce model accuracy and when to favor flexible output formats over standard serialization. Then explore "Modal: ML Infra That Does Not Suck" with Developer Advocate Charles Frye, who presents Modal as a serverless Python infrastructure solution for building cloud applications with GPUs and foundation models. The 68-minute session features these industry experts sharing practical approaches to overcome real-world AI implementation challenges.