Mr. Vinay Soni

Building Scalable AI Systems with Feature Platforms

Abstract:

As machine learning systems scale, the complexity of managing features becomes one of the biggest barriers to reliable, reusable, and production-grade AI. Feature Platforms have emerged as a foundational solution—offering centralized infrastructure to define, compute, store, and serve features across training and inference environments.

In this keynote, we’ll explore how Feature Platforms accelerate model development, reduce operational debt, and enforce governance at scale. We’ll cover core architectural patterns, design principles, and best practices, drawing from real-world implementations. We’ll also discuss the emerging future of these platforms.

Whether you’re an ML engineer, platform architect, or data leader, this session will equip you with a strategic understanding of how Feature Platforms can serve as the backbone of scalable, production-grade AI systems.

Profile:

Vinay Soni is a seasoned software engineer with deep expertise in machine learning platforms, cloud computing (Azure and AWS), and large-scale distributed systems. Currently a Staff Engineer at LinkedIn, he leads backend efforts for LinkedIn’s Feed recommendation system. Previously, he led critical platform initiatives at Apple, Amazon and Microsoft, building scalable systems for data transformation, control planes, and ML infrastructure. Vinay holds an MS in Computer Science from Georgia Tech and is skilled in technologies like Java, Python, Distributed systems, Machine Learning, Apache Spark, Kubernetes, and more. He is passionate about building robust, high-impact systems.