Speed and Scale AI Inference Operations across Multiple Architectures

When executing inference operations, developers need an efficient way to integrate components that deliver great performance at scale while providing a simple interface between the application and execution engine.

So far, TensorFlow* Serving has been the serving system of choice. But, with it come challenges including its lack of cross-architecture inference execution on GPUs, VPUs, and FPGAs.

The 2021.1 release of the Intel® Distribution of OpenVINO™ toolkit solves these challenges with its improved Model Server, a Docker* container capable of hosting machine learning models for high-performance inference.

Join principal engineer and AI solution architect, Adam Marek and AI developer tools product manager, Zoe Cayetano to learn about this serving system for production environments, including how to:

• More easily deploy new algorithms and AI experiments for your AI models
• Take advantage of a write-once, deploy-anywhere programming paradigm, from edge to cloud
• Use Docker containers to simplify the integration of AI inference with a wide range of platforms and solutions

• Download the Intel Distribution of OpenVINO toolkit—includes nearly 20 developer tools and libraries for creating cross-architecture applications.
• Sign up for an Intel® Developer Cloud for oneAPI account—a free development sandbox with access to the latest Intel hardware and oneAPI software.

Adam Marek
Principal engineer and AI solutions architect, Intel Corporation

Adam focuses on the distributed platforms and components for deep learning platforms. He joined Intel in 2003, but his expertise in software architecture and design has been honed since 1998. He has worked for companies including Softmatic and Oke Software and Communication on projects for telecommunications, embedded speech processing solutions, and large-scale back ends for secure online services. Adam holds a master of science degree in computer science from Gdańsk University of Technology and a number of patents in various software fields.

Zoe Cayetano
AI developer tools product manager, Intel Corporation

Passionate about democratizing technology access for everyone and working on projects with outsized impact on the world, Zoe is a product manager for AI and IoT working on various interdisciplinary business and engineering problems. Prior to Intel, she was a data science researcher for a particle accelerator at Arizona State University, where she analyzed electron beam dynamics of novel X-ray lasers that were used for crystallography, quantum materials, and bioimaging. She holds bachelor’s degrees in applied physics and business.

Dariusz Trawinski
Deep learning software engineer, Intel Corporation

Dariusz specializes in deep learning applications and solutions. Joining Intel in 2000, he has honed his software expertise through web application development, system administration, information security, and data center and cloud management. Most recently, his focus has been on optimizing AI application performance and improving user experiences in deep learning and inference platforms. Dariusz has a master of science degree in telecommunications from Technical University in Gdansk, Poland.