Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Discover the practical realities of deploying generative AI at edge devices in this 44-minute conference talk featuring Yuriy Krustalev from Liquid.AI, an MIT-born startup. Learn how to overcome the extreme constraints of edge computing environments where memory limitations, connectivity issues, and precision requirements create unique deployment challenges. Explore Liquid.AI's approach to using ExecuTorch for bridging Python-based training with efficient C++ deployment, while mastering essential techniques including smart memory caching, quantization strategies, and kernel optimizations that enable compact models to deliver powerful performance. Understand the critical importance of building modular, cross-platform deployment pipelines that work seamlessly across Linux, Android, iOS, Windows, and other operating systems. Gain insights into why interface design through C, C++, CLI, and HTTP protocols is equally crucial as inference optimization for successful edge AI implementation. Perfect for AI developers, edge hardware vendors, and professionals designing next-generation inference systems who need practical guidance on production-ready edge AI deployment.
