Machine learning engineer Daniel Bourke completes Google's Cloud Platform Big Data and Machine Learning Fundamentals course in one day, working through Compute Engine, Cloud Storage, TensorFlow, and the Vision API. He emphasizes the value of cloud-based ML pipelines for production and pre-trained Google APIs, mixing study sessions with gym breaks and a personal banana classification project.
Chapters
Gonna need some coffee for today. It's a big day on tour. First of all, got the Google Cloud Platform training that I'm supposed to be doing, and then after that, do that for a little bit. We've got a talk with Shaikh and Ahmad. The magic of the internet. Okay, so let me start recording. Yeah. And then probably do a workout with Josh, and by the end of today, try and figure out whether I should continue on with the data science assignment 4 or keep going with the little project we were working on yesterday. Yesterday, banana, not banana. Temperature. He's not a banana! At the moment I'm more excited by the project I'm— my own project. You know, when you're really interested in something. I mean, I know I need to do the data science assignment, but you gotta chase what you're excited by, right? Let's do it. Morning, bro. What are you up to today? Me? No, the other person in the room. Oh, go for coffee, I think. Go for coffee with the rent? Beautiful, good idea. So the Google Cloud Platform certification is a data engineering course of some sort, so I can figure out how to use the Google Cloud Platform, like AWS, like Microsoft Azure, to host data engineering pipelines. So that might be a machine learning model or something like that, because It's a part of what we're working on at Max Kelsen. We partnered with Google Cloud and they've given us some free courses to do on Coursera, which I'm really thankful for. I'm really excited about because if we need to run data science pipelines or machine learning models, whatnot, and get them into production rather than just having them sit on the cloud, we want people to be able to use them. So that's one really valuable thing about getting a product to market is having it accessible to to people all over the world, and Google Cloud Platform allows for that. I mean, they all do, but I'm going to be studying Google Cloud Platform today. I own the machine and manage the software on it. I installed everything. I had HootHacks on that machine. It was on the— Alexa, stop the alarm. It's a real cool module already. And where's the value in learning stuff from here? Well, I think one of the use cases that, that I think his name is, is Tuck or Fluck? Luck. Luck, who's an awesome instructor. He said when he was an academic, he used to run machine learning models on weather and whatnot, and he used to run all the computing on the university's on-premise. And, and now when he switched to a startup, they used the cloud and they could run the same thing And this is, this is, I mean, if you're, if you're new to machine learning or new to data science or new to any cloud computing, you might be thinking, well, the cloud is the only way to go. And yeah, I agree with you. But in the past, he said what used to take them 4 years to run, they were running every 2 weeks at startups and using the cloud because that's the access we have today, right? In the past, you might have had to have your own machine, have a have a really big, big piece of hardware sitting in your room before you could even run something. But now we can use this baby little laptop, dial up to the cloud and run it on some awesome computing hardware that we don't even own and then pay per second. And so where am I up to now? Well, the next is GPUs on demand. That's a lecture that we're up to. And I'll run through some of the curriculum in a sec after I've talked with my boys. So we'll do that and then we'll get back into this course. Dialing in now. Yo, what's going on, team? So I want to start off with some brief introductions, and I'm hoping that both of them can tell me about where they're from, are you employed or in school, and what you're currently interested in right now. Daniel first, yes. Ladies first. Ladies first. Alright, thanks, man. Hello everyone, nice to meet you. My name is Daniel Bourke and I am currently a machine learning engineer at Max Kelsen, which is a technology company. We have about 25-27 people in Brisbane. I like it. I like it because it's small, but we work on big projects. I'm also a student because I think studying is a lifelong journey and it's all about just continual learning. And at the moment, I always see myself as a a beginner with whatever I do. So, no matter how much experience I have in something, there's always something more to learn, which is exciting. And that was really fun. You'll probably be able to see the full video on Hamad's or Shaikh's channel or something like that, but I'll link it in the description once it's done. Really good conversation about how to get started in AI. Data science. Shaikh's 14— 15, sorry, from India and learning AI. It's, it's a fun place to be right now. Back to, back to study. All right, my Pomodoro timer is telling me it's time to take a break, so let's start that. I'll go for a little bit of a walk, shall we? It's been great fun so far. Been through 2 labs, and the first one was, was Compute Engine. So getting Google's Compute Engine— I really want to sit down, actually. I've been sitting down for the past hour. So first one was Google's Compute Engine, and the second one was Google's Storage, or GCS, Google Cloud Storage. And why is that? Why are they important? Well, that's what every computer is made of, right? It's It's made of something that can do compute, CPU, and something that can store things, RAM and hard drives. And so what essentially the labs will go through is just showing you how to get one started. And I mean, it's a few clicks of a button, and that's what's important. That's why I'm all about things that are simple. And if something is too hard, takes too many steps to use, then I'm not gonna use it. So I think Google making their cloud being easy to use is a big plus in my book. So the next time I want to install something, it's as easy as git clone, just like you would on your local machine, but you can do it on the Google Cloud Platform. And I need to take care of these puppies. What you barking at? What's going on? And then we'll get back to some study. Break time's over, back to it. These puppies, come on! Break time. We're going to finish all of these modules today. Trust me. All right, time for another break. I think my brother's waiting to go. Hey bro, what are you doing there? Waiting. What do you want to do? What have we got there? You got the drift trikes ready? Yeah, I'm waiting. Collision. Course. It's like bumper cars and drift trikes in one. Not a bad way to spend a study break, hey bro? That wasn't bad. How's your course going? Yeah, good. Um, just about to wrap it up actually, almost finished, about to move on to the next one. Yeah, Josh is doing iOS, doing Angela's iOS. Angela Yu, yeah, yeah. She's a good teacher, hey? Yeah, yeah. If you want to learn iOS, jump onto that course. But now it's time to back to the Back to the Google Cloud. Oops, sorry man. Check it out, we've, we've worked through a large portion of this course. What's next? Next is carrying out machine learning with TensorFlow, and I just learned about Datalab, which you may have been— you may have used it before, but Datalab is is a Jupyter Notebook that runs on Google's cloud platform and allows you direct access to their compute power. So that would come in pretty handy. We're gonna go to gym now and then come home and finish the rest of it. Who's keen to lift? Xeno, you in? You coming gym? Yeah, I'm coming gym. Josh, you coming gym? Yeah, you look keen for it, bro. All showered up, finished gym. How cool are my slippers and PJs? Now, what is it? It's Friday night. I think it's like 9 PM, 9:37 PM. We're going to finish this course tonight. I've got probably an hour worth of study to go, but I promised myself I'm going to finish it and I don't break promises with myself. So let's get into it. I'm up to Machine Learning with TensorFlow on Google Cloud. The next thing that we look at, and I think of this whole idea of doing the things that transformational— We did it! Check it out. Congratulations, you've successfully completed Course 1 of 5, Google Cloud Platform Big Data and Machine Learning Fundamentals. Oh yeah, the last course, or the last part of it, was my favorite. It was the ML APIs and what, what could you use those for? So they're pre-trained Google's machine learning models on a bunch of Google data. So, for the problem we were doing yesterday, Josh, remember banana, not banana? We created our own model, but if we wanted to use one of Google's pre-trained ones, say on the Vision API, we could use that. And check it out, they give demos of it in a Jupyter style notebook. So, here's a Vision API for translation. It will read that sign and then give you the output here. Boom. And then it will translate it for you. Translate it. Please care for and protect the health to create a beautiful water environment. What a lovely sign. And then we go down. Oh, here's the Vision API. So if you go to cloud.google.com/vision, you can test out the API. And now we're going to see how it, how it works. You can drag and drop an image. Here's our banana from yesterday. How's the Google Cloud API going to go? Oh, much better. Banana detected 87%. So maybe, maybe we'll look into that to see if it can do, do more than one thing. But it's definitely past my bedtime now. It's, it's 10:57 PM. It's time to rock and roll. We got the whole course done in a day. That's not even going to focus, but there we go. All All done, all ticked off, can go off the notebook. Boom. There we go. Done. I learnt a new saying today. I think it's not my first language, so forgive me if I get it wrong, but my friends Hamad and Shaikh taught it to me. It was "pana ban karo." Now, I really hope that's the right sign-off message, but even if it's not, even if it's saying it wrong, It's not. Wow, see how tired I am? It's been a good day. Pana, bandh karo. Peace out.
