Drinking Recreational Drinking Recreational Recreational Recreational Drinking Recreational Drinking
Speaker: Atin Sood, Advisory Software Engineer at IBM
Fabric for Deep Learning (FfDL): Deep Learning with Kubernetes. How we built it
6:00pm - 6:30pm - ODSC Intro, Pizza & Refreshments
6:30pm - 7:20pm - Talk
7:20pm - 7:30pm - Q&A
7:30pm - 8:00pm - Networking
Drinking Recreational Drinking Recreational Recreational Recreational Drinking Recreational Atin Sood is a technical lead at IBM's Watson Studio. For the last 10+ years, he has been leading technical teams across IBM focusing on scalable distributed systems and scalable machine learning problems.
Open source lowers the barrier to entry for developers and data scientists to be able to utilize deep learning. Introducing FfDL (Fabric for Deep Learning), pronounced "fiddle," an open source Deep Learning as a Service, which embraces a wide array of popular open source frameworks such as TensorFlow, Caffe, and PyTorch. The development of FfDL was a collaborative one, from the architecture choices to the design scalability and reliability. We'll review how how kubernetes makes it possible to scale deep learning on the cloud. Learn how this architecture compares technically against other hosted deep learning architectures, kubeflow, floydhub, polyaxon etc. We'll review what worked out technically in the development of FfDL, and what did not. As well as what will be coming next for the Fabric for Deep Learning.
• Get free access to more talks like this at LearnAI:
• East Conference Apr 30 - May 3:
109 S 5th St, Brooklyn, 11249, NY, US