Hemang Shah

Vector Search in IP Protection

Abstract:

The talk covers the problem of intellectual property protection and how computer vision helps tackle it in the industry. It goes over the different types of intellectual property, how big is the IP infringement problem, what are the different families of image models that can help tackle different parts of the problem, the role vector search plays in this domain, different vector search algorithms and their applications.

Profile:

Senior Software Engineer with 5+ years of experience architecting scalable ML-powered systems. Currently part of Zoom’s AI org developing an enterprise workplace suite, where I design search and agentic retrieval systems that connect enterprise data to AI-driven workflows. Previously at Amazon, led high-impact ML systems for IP protection and search at massive scale, delivering measurable gains in operational efficiency, associate productivity, and risk reduction.