Presentation: Actors for Agents: Patterns for Production AI

Manju Rajashekhar: I'm Manju Rajashekhar. I started my career more than a decade ago at Twitter where I helped solve the fail-mail problem. At that time, I worked on distributed caching and I was also a prolific open-source contributor around that time. A lot of the software that I wrote and open sourced around that time is still used today at companies like Uber and Wikipedia to scale your caching systems. GCP and Azure documents it as a reference implementation for designing your caching systems for scale. After Twitter, I ended up founding a company in the deep learning and search space. Our product was search as a service and recommendation as a service. That company was acquired by Etsy, where I spent eight years as a VP of engineering. I ran all of their AI and machine learning functions that included search, ads, recommendation, and personalization engine. Most recently, I started a company called Mad Labs with a couple of really wonderful colleagues of mine. We are an agentic platform company focused on building agents that have reliable execution over long horizons and they're very token efficient and have a strong governance structure around it. Today, I'm really excited…