Staff Software Engineer at LinkedIn, Seattle, Washington, USA

Title of the Talk :
From Code to Scale: Designing Backend Systems for AI-First Products

Abstract of the Talk:
As AI transitions from a research tool to the engine behind modern digital experiences, the role of backend engineering has evolved. AI-first products—like feeds, recommendation engines, and personalized assistants, demand systems that are not just fast and reliable, but also aware of models, data drift, and user feedback loops.

This keynote explores how to design and scale backend systems purpose-built for AI-centric applications. We’ll examine how traditional API boundaries shift in the presence of models, why observability needs to extend to features and inferences, and how fallbacks and resilience strategies evolve in probabilistic systems. Attendees will walk through design patterns for inference APIs, feature stores, model versioning, and failure handling, using real-world examples from recommender systems and user-facing AI products.

Whether you’re scaling to millions or building your first intelligent feature, this session will equip you with the systems thinking and architectural patterns needed to bridge the gap between data science and production-grade infrastructure.

.