Register
While LinkedIn overflows with thought leaders declaring every new research paper a 'game changer,' few are sharing real experiences of building and scaling AI products in production. Over twelve months, Phil and his co-founder scaled an AI assistant from prototype to serving 10,000 users, outperforming even deep-pocketed incumbents like Salesforce and Slack AI.
Working at the frontier of AI systems meant learning through trial and error. They had to discover which traditional architectural patterns truly work with AI systems. More importantly, they had to identify which ones fundamentally break down when faced with stochastic behavior and stateful requirements.
In this talk, Phil shares practical insights from this journey, including how traditional approaches like microservices needed radical adaptation for AI systems, and which novel patterns emerged as essential for production success. Expect concrete architectural lessons drawn from real experience - patterns that actually work when building AI products that ship.
Speaker
Phil Calçado
Founder & CEO @Outropy | Previously Pioneered Microservices Architecture @SoundCloud | Scaled Systems @DigitalOcean and @SeatGeek | 20+ Years Experience in Software Development
Phil Calçado is the founder of Outropy, the AI developer platform for software engineers. Previously, he pioneered microservices architecture at SoundCloud during their growth from millions to hundreds of millions of users. His experience scaling systems with small teams spans roles at DigitalOcean and SeatGeek, consistently helping underdogs compete with tech giants. He lives in Park Slope, Brooklyn, where he divides his time between building AI products and working on his mediocre Brazilian Jiu-Jitsu.
