On the 31st of March, I attended one of my first AWS events: the AWS Summit in Brussels.

When entering the venue, my first impressions weren’t as expected. I envisioned the location to be one with only a limited amount of rooms to hold presentations. Instead, it turned out to be a huge event hall with multiple stage areas. Furthermore, the welcoming was tremendous, and everyone was super friendly and excited about the event and the speakers.

I went exploring and stumbled upon some familiar faces, and I met some colleagues of Cloudar (a company with which we work together closely). They told me a lot about other AWS events and this one. They gave me helpful tips and led me to exciting sessions. Most of the sessions held during the event were based on AI, data and AWS GreenOps.

Cloudar booth

AWS Control Tower

During the AWS Summit in Brussels, I had the opportunity to learn more about AWS Control Tower.

Control Tower is a service that allows you to set up a multi-account AWS environment that has already been made secure and comes with out-of-the-box best practices. It creates a free-of-charge environment. However, you are paying for the resources spun up in the background.

Another nifty feature of the AWS Control Tower is that it allows you to apply guardrails. Guardrails can be seen as strongly recommended high-level rules. These rules help enforce policies and detect policy violations. Furthermore, Control Tower will give you an overview of how each account conforms to your policies.

AWS Deepracer

I also had the opportunity to deep dive into machine learning with AWS Deepracer. Deepracer can be seen as an easy-to-step-into machine learning technique. It allows you to provide code for an autonomous car. This code is called a model. A model contains information and rewards specifically for your vehicle. The preferred code to use for creating a model is Python.

A model can then be used virtually to train your car on a specific track. Furthermore, you can reward the model whenever it does a particular action, for example:
If you want your car to be driving in the middle of the track, you can reward it with higher marks than when it drives away from the centre of the track. You can also give negative awards.
Let’s say you don’t want your car to be driving off track, and then you’ll have to use the boolean off_track. If set to true, a negative score is awarded. This score will eventually help your car learn a specific behaviour.

The goal of creating such a model is that you eventually upload it to a real-life autonomous car. These cars will be available to drive during any AWS event on a life-size circuit. The life-size circuit could be the one you trained your model on, or it could be a completely different one.

If you ever get the opportunity to create your own model, be sure to do so. It is a great way to compete with colleagues as well as if you are any good, you can win prizes and attend real-life tournaments.

End of the Day

At the end of the day, I went home with some goodies (you can never have enough goodies!) It was a long day, and I still had a long drive home. While driving, I looked back on my first experience of a work-based event. I had an extraordinary time, met some interesting people, and learned new things.
In the end, I am super thankful that I was allowed to attend this event. Up to the next one!