Hosted at Navy Pier, the Google Cloud Summit offered a great opportunity to catch up, connect, and learn about the new cloud services offered by Google Cloud Platform. The Academic Software Development team (Rodolfo Vieira and Alex Miner) attended the event.
In his “Founder’s Letter” from 2016, after taking the role of CEO at Google, Sundar Pichai announced the move from a “mobile-first” to an “AI-first” company. Looking at the agenda for the Google Cloud Summit, it is easy to understand that the message was not taken lightly at Google.
The agenda offered a glimpse at the latest services related to machine learning ranging from the AutoML to orchestration complex ML pipelines. AutoML is particularly interesting as it allows developers with green/limited experience in ML to train quality models and use them in production environments. The set of services available under AutoML comprise beta versions of the Vision, Natural Language Processing, and Translations.
Cloud Data Labs is a wonderful solution for data exploration (built on Jupyter notebooks) that provides [limited and free] compute power for training custom ML models as a proof-of-concept, and then by taking advantage of its deep integration with the Google Cloud Platform, it can spawn up and handle the orchestration required to run the training jobs on large datasets at scale.
Kubernetes (GKE) is a known entity for container orchestrators and, along with Docker Clusters have captured the imagination and mindshare of DevOps around the world. KubeFlow, provides a toolkit for running ML jobs in Kubernetes clusters.
Other highlights included the concept of “Cloud Worker” — Forrester’s research highlights that currently 1 in every 4 workers in today’s enterprises are already cloud workers by using cloud native applications for an average of 4.6 hours per day.
PS: We are hiring.