As with every re:Invent, AWS has launched a number of great services and features. One we’re excited about is the addition of a new AWS Certification for Machine Learning. Here’s a bit about my own journey to get the Machine Learning Certification at re:Invent.
A brief history of AWS certifications
AWS released 8 certifications so far (9 if you dare to count the cloud practitioner). Before the specialties, AWS used to offer 5 role-based certifications: 3 Associates (Developer, Sysops Admin, Architect) and 2 Professionals (Devops, Architect). In 2017, AWS released 2 more specialized certifications : Advanced Networking and Big Data. In 2018, they finally added the Security Specialty.
Yet even if the Big Data certification covers some machine learning aspect, something was missing.
And here we are…
In October 2018, the AWS Machine Learning certification was announced, in beta version. My first move was to register, for re:Invent 2018, in Las Vegas.
It was a step in the unknown as :
1 – No training or feedback of any kind were available.
2 – It was my first re:Invent ever!
A brave new world of intelligence
A world with AI is a world of intelligence with a lot of data, but paradoxically not so much information. They say in 2020 there should be around 50 ZettaBytes of data in the cloud. Knowing that zettabytes is one million petabytes, let’s just say it’s big. From that data, not only can we extract precious information about the past and the present with big data technologies, but now we can predict what the future will look like thanks to Machine Learning technologies.
Customers want those predictions badly, and AWS knows it. You can tell just by looking at the announcements; many are related to Machine Learning like AWS Forecast. AWS has even created a new marketplace dedicated to machine learning!
No need to use a ML model to predict that those kind of services will change the face of the cloud computing very, very soon.
Machine Learning is Everywhere
Last year, AWS had already put some magic in their DeepLens Camera. Recently they used an interesting kind of machine learning called reinforcement learning for their new product : the DeepRacer, an intelligent self-driving toy car.
(More info is available here : https://aws.amazon.com/deepracer/)
Curious about what a DeepRacer could do, some of the Onica team participated in a DeepRacer workshop. A few lines of python code later , our team was the first to cross the virtual finish line (simulated through robomaker), and we won two beautiful DeepRacers.
re:Invent 2018 was starting well. Let’s step deeper into our re:Invent certification journey.
Being certified at re:Invent comes with some advantages. One of them is being able to access the certification lounge. It’s a peaceful place where you can eat, drink, and get some rest when you need a break from the crowded re:Invent jungle (we are talking about 60,000 people here).
Believe me, it was very welcome sometimes!
When you register for the first time, they give you an AWS Certified sticker for your attendee badge so you can come and go whenever you want.
A cool thing at re:Invent is they give you swag, a lot of swag.
And it gets cooler when you are certified. You get even more goodies: a t-shirt, a nice cap, an useful bottle and many pins.
AWS booth : Share your success
Certifications seem to be very important for AWS. It’s a way to demonstrate your abilities with AWS services. In the certification lounge, they ask you if you want to share your story with others.
Inside the booth, you can pick a question you will answer in a short video. This video can end up on their website or even on Youtube with your consent.
Hopefully we will inspire somebody!
AWS Certification Center
One of the AWS certification centers was at the Venetian hotel. I sat for the exam on Friday morning, 8AM.
It happened to be a big re:Invent beginner mistake since re:Play was on Thursday night.
When you register for an AWS certification, you acknowledge that you can’t go into details with the questions. So even if I stay purposely vague about the exam, I can still tell you this :
It’s a machine learning exam! And I was very surprised how some of the questions were not AWS related. I would say 50 percent of the questions were about machine learning itself (frameworks, algorithms etc..). The other 50 percent could be divided into 2 groups :
- 20 percent of ‘Big Data like’ certification. A lot of questions about what kind of storage to use given a specific use case (S3, EBS etc..). The kind of questions I had during the Big Data certification test.
- The remaining 30 percent were mainly about SageMaker. I had maybe one or two questions about others AWS services like Polly or Rekognition.
No Reward, No Swag
My biggest disappointment! Everyone who successfully passed the exam were given a lot of nice goodies. When it came to my turn, as I sat for a beta exam (no live result), they did not give me a thing. I tried everything, even my saddest look: nothing for beta exams!
I will have my result in 13 weeks only…we’ll update you if I pass!