Speaker "Michael Zimmerman" Details Back



Virtualizing ML/AI and data science workloads


ML/AI infrastructure today is monolithic and mainly designed around generic servers (some are GPUs). We anticipate that soon, a much wider variety of accelerated compute servers will be offered - AI ASICs, variety of GPUs and even newer CPU architecture designed for neural networks. The focus of this talk will be how a modern heterogeneous data center will offer elastic instances and virtualization for the ML/AI and data science workload. In specific, how the the new accelerated hardware will not dictate the applications and data placement, and how we will be able to run the ML/AI workloads on any hardware instantaneously without being limited to the scale and location of the hardware.
Who is this presentation for?

Prerequisite knowledge:

What you'll learn?


Michael has more than 20 years of experience in building ML/AI platforms, compute and storage infrastructure, and large-scale networking data centers. Michael served as CEO and VP products in four start-ups which were acquired by industry leaders (AWS, VMware, Mellanox). Recently Michael was the CEO of Bitfusion, the elastic ML/AI platform which was acquired by VMware, to serve as the lead technology for ML/AI virtualization in hybrid cloud. Until Sep-2017, Michael was the infrastructure GM/VP at Marvell, leading the transformation to a $600M+ leading position in enterprise and data centers. Before Marvell, Michael was an executive for private companies considered as industry leaders such as Annapurna Labs, who develops high-performance distributed storage (acquired by Amazon Web Services), and Tilera, the MIT-based mass-compute company, which was acquired by Mellanox. Between 2001-2003 Michael worked in Radlan, an Israeli start-up, which developed a hardware-agnostic networking operating system used by several tier-1 players as a standard network fabric. In 2003 Marvell acquired Radlan, and Michael served in a range of management positions including compute and networking VP/GM. Michael is a Stanford graduate where he attended the Executive Business Program, as well as, holds an MS in Computer Science from NSU University, an MBA from Tel Aviv University and a BS in Electrical Engineering at the Tel Aviv University, where he earned Summa Cum Laude.