Ram Iyengar
Ram Iyengar is an engineer by practice and an educator at heart. He was (cf) pushed into technology evangelism along his journey as a developer and hasn’t looked back since! He enjoys helping engineering teams around the world discover new and creative ways to work. He is a proponent of product development and engineering teams that put the community first.
Session
What does operational overhead look like in the era of MLOps? If you're grappling with this question, like many others, and would like a way to apply the paradigm of containers and cloud native to AI workloads ― you're in luck.
There is an effort underway to align AI workloads with the knowledge we have of operational excellence in cloud native. The CNCF Sandbox project ModelSpec brings much needed clarity to MLOps workflows. It provides the right abstraction to be able to define how DevOps and cloud native practices can be applied for machine learning operations.
APplying the ModelSpec is the KitOps tool. It helps bridge the gaps that currently exist in the tooling space for MLOps. It creates a "Docker"-like interface for AI workloads and makes it easy and efficient to work with models on Kubernetes (or other container runtimes).
In this talk, I aim to bring together the ML overhead, how cloud native paradigms can help, the ModelSpec, and KitOps. Together, all these will help expose an important painpoint in productionalizing AI in the workplace. Let's eliminate all the disconnected ways in which data teams, developers, and operations folks are working by using the principles that will be highlighted during this talk.