Mastering Cloud: can we do better than certificates?

Nowadays, cloud computing is ubiquitous. For most companies, the advantages are clear: you can scale your infrastructure elastically instead of having to over-provision; you only pay for what you use. It allows companies to focus on their core competencies which accelerate innovation. Startups get to market more quickly, and can land those first clients without first having to burn through a pile of cash for months; established players iterate more quickly, and with their nimbleness and agility one-up the competition.

In short: the Cloud is where you wanna be. It’s where the cool kids are. But how to get there? Take for example Amazon Web Services (AWS), the largest of the three most popular cloud providers (AWS, Azure, and Google Cloud Platform). With more than 200 services, AWS can feel overwhelming for people new to the cloud scene. Where do you even start? And the list of services only grows longer and longer, year after year.

For many, a good place to start is cloud certificates. A certificate is typically geared towards a specific role or profile in the data space, such as BI analyst, architect, data engineer, etc. It introduces you to a subset of services deemed relevant to that role. There are a few notoriously difficult certificate exams that require very detailed and specialised knowledge, but most of them only expect you to have a bird’s eye view of the cloud landscape. To know "the lay of the land", if you will. More importantly, however, is that for a lot of cloud novices, certificates can help you get a foot in the door. Landing a job without a university or college degree is more commonplace than a couple of years ago*, and obtaining a cloud certificate goes a long way in convincing a potential employer of your potential. But what to do once you’re in?

You see, here’s the main gripe I have with cloud certificates: the knowledge you gain from studying for a certificate exam is passive and often does not translate well into practice. In a way, it’s a clever scam that exploits our Darwinian urge for self-improvement while acting as an advertising channel for the larger cloud providers. Advertisements paid for by you.

Let me explain a bit more in detail what I mean by that.

The most popular way to study for a certificate exam, is via e-learning resources. As a consumer of the content, it’s rather easy. You can watch the material from the confines of your comfy home. You’re learning something that might benefit you in the future. It increases your employability. It makes you feel good about yourself**.

It’s also not too difficult to game this certificate system. There are several websites where you can buy example exam questions for a "small" fee ("small" as in "small in comparison to the price of the actual exam"). You can pass many of the entry-level exams by iterating on these questions, until you know the answers by heart.

But in the end, how much have you actually learned? You now know the latest and greatest Cloud provider X has to offer (during an exam, when in doubt: their newest, fully managed, Serverless™ service Y is probably the correct answer). You might even remember that an AWS Lambda function times out after 15 minutes and impress your colleagues during standup with this and other bits of trivia. You’re familiar with the basic building blocks, but can you now build a castle? An entire city?

I think it’s fair to say that there is a significant gap between what you learn from such courses, and what you need to know to perform well in your job as a platform or data engineer. Most people learn their chops on the job. And those first few months can be quite stressful. It easily takes several months, up to a year before you start feeling a bit more confident in your new role. Some people might argue that is the only way you really learn; that this is the harsh reality we live in, and that we got to suck it up. If you’re lucky (like I was), you’re assigned a mentor or are part of a team that supports you along the way. If not, well, you’ll probably get thrown into the deep, and have to struggle to keep your head above the water. Which is neither sustainable nor is it equitable for those who cannot afford to spend their after-working hours trying to catch up. Can’t we do better?

At Data Minded, we believe there are better ways to bridge that gap. In fact, we have found an entire Academy based on this belief and a few core educational principles. For most of our courses, we make use of a flipped classroom approach. Instead of passively listening to a teacher for two hours, bravely trying to shovel a bunch of theoretical concepts into your brain, we focus on actively applying those concepts in practice. The theory sessions are prerecorded, and you can watch the videos at your own pace. During the interactive sessions (either in person or remotely), you are in the driver's seat. Our instructors are there to support you and help you navigate your way toward a solution. In essence, you could call it an Agile approach to learning: we aim at creating a short feedback loop, by offering small, bite-sized chunks of feedback; the real skill is in learning to ask the right questions that help you get unstuck.

In our Introduction to Cloud Providers (with AWS) course, we build an actual mini-platform for an imaginary data science startup. Although the cloud provider used in the course is AWS, we try to keep it as generic as possible. We’re not married to a single cloud. Our architecture includes all of the different components of a mature data platform: computing, storage, networking, and access control. These concepts are, indeed, transferable between clouds.

In addition, we discuss the flaws and shortcomings of our construction and the way we built it. Clicking things together in a graphical UI is great for learning purposes, but does not scale very well. Without going into too much detail, our goal is to convince you that in the mid-to-long term it pays dividends to focus on automation, and the power of infrastructure-as-code tools.

Needless to say, what we build in our half-day training is just a toy architecture. Rome was also not built in a day, let alone an afternoon. There is nothing that beats actual on-the-job experience. Learning about the intricacies of a system takes time, and typically involves getting bitten quite a few times. Still, I believe our flipped classroom approach is a considerable improvement compared to the traditional, certificate-oriented path that most people take when it comes to learning about Cloud.

If you’re interested to learn more about building stuff in the Cloud, sign up for the next edition of our Summer School! Or check out our other courses, at www.dataminded.academy/courses.

*Google now hires people without a CS degree
**An interesting read on this subject is the article Productivity porn by Caleb Schoepp