The Developer Show (Google Cloud Platform Podcast: Google Cloud ML)

The Developer Show (Google Cloud Platform Podcast: Google Cloud ML)

TIMOTHY JORDAN: Hey, everybody! I’m here at the Google Launchpad
Space in San Francisco. Now, one of the cool things
about a space like this is all the great people
that you get to run into. For example, the crew that
does the Google Cloud Platform Podcast is recording
in this room right now. And I kind of want
to crash the party. Want to come with me? TIMOTHY JORDAN: Hey, guys! BOTH: Hey! – How you doing? – You recording a podcast
or something in here? – Yeah, kind of. – You mind if I crash the party? – Please join us. – Hi, everybody! I’m Timothy Jordan. – Yeah, Timothy Jordan
from the Developer Show just crashing our podcast. – It’s what we do. – Yeah! How are you doing? – How am I doing? Great! I hear that you guys have
a question of the week, and I would like to
ask that question. May I? – Please. Please go right ahead right. – OK, so, on the Developer Show,
we’ve been talking a lot about machine-learning recently. And a lot developers
know that TensorFlow is a fast, flexible, and
scalable open-source source machine learning framework. But what they don’t
all seem to know is that you can train
and run TensorFlow models on Google Cloud Platform. Can you tell us a little
bit more about the details? – Yeah, that’s a great question. So I think that what
you are referring to is what we call the Google Cloud
Machine Learning Platform– Cloud ML for short–
and, basically, it allows you to train models
on Google Cloud Platform infrastructure. Yeah. – And that’s, I
think, really key, because once you get into
machine learning pretty deeply, your data is so
large and training takes so long on
your desktop at home that having a scalable
infrastructure to train on becomes critical. – Yeah, it is definitely
very, very important if you’re going to
start training models. So you can start with
either your own model or choose one from the
pencil flow framework, which is open source. And then you’re going to
need to train those models. And no matter how big
the data you’re using is, you’re going to be able to do
that on Manage and completely scalable infrastructure
on Google Cloud Platform. And then once
you’re done, you’re going to be able to
serve those train models, so you can do predictions based
on that with live traffic– again, on Google Cloud ML. – Awesome. – It’s also worth
noting, though, if machine learning isn’t
particularly your thing, we have pre-baked
services that you can use. So we have the Vision API
Speech, natural language processing, and the
Translation API. These are REST APIs
that have already had all the machine-learning
training, and research, all done for you. And they’re a REST endpoint. So if you can call a REST
endpoint from any language, you can use those things. – Well, I will consider
my question answered. Thank you, gentlemen. – Awesome. Thank you. – Thank you. – I’ll let you get back to your
regularly scheduled podcast. – See you! – See you, guys! [MUSIC PLAYING] – Make sure to tune in
weekly for new episodes to teach you all about
Google Cloud Platform, and how you can get
the most out of it. – And check out all of
our previous episodes on, and
make sure to subscribe.

6 thoughts on “The Developer Show (Google Cloud Platform Podcast: Google Cloud ML)

  1. I tried to use GCP. And i feel it was developed by Google to use by them selves. it is not possible to understand the logic of UI. In Azure i was needed 10 minutes to set up my real node app. I GCP i spend 2 days and im done! Will never use or recomend Google cloud for anybody.

Leave a Reply

Your email address will not be published. Required fields are marked *