AWS Deep Racer


Start course

This is a short refresher of the 7 AWS machine learning services announced at Re:invent 2018 which will cover:

  • Amazon SageMaker Ground Truth
  • Amazon Forecast
  • Amazon Comprehend Medical
  • Amazon Textract
  • Amazon Personalize
  • Amazon SageMaker RL
  • AWS DeepRacer

Learning Objective

  • It aims to provide an awareness of what each of the ML services is used for and the benefit that they can bring to you within your organization

Intended Audience

  • This course would be beneficial to anyone who is responsible for implementing, managing, and securing machine learning services within AWS


  • You should have a basic understanding of Machine learning concepts and principles to help you understand how each of these services fit into the AWS landscape

Related Training Content

Introduction to Machine Learning on AWS

Applying Machine Learning and AI Services on AWS

AWS Machine Learning - Specialty Certification Preparation


Have you ever felt like you wanted to get into machine learning, but just couldn't make heads or tails or where to start and your attention span for all the abstract terminology is in the microsecond range? Or maybe creating a machine learning model that makes inferences for a fictitious scenario just doesn't cut it in terms of interest factor. Maybe you're just keen to experiment with reinforcement learning tech in a global community context where you can share ideas. If you answered "yes" to any of these questions, then maybe you should try our AWS DeepRacer. 

To get the next wave of machine learning enthusiasts on board and familiarized with AWS machine learning technologies, AWS has released AWS DeepRacer. AWS DeepRacer is a new global racing league for autonomous handheld sized racing cars. The idea is quite simple, build and train a reinforcement learning model that can be uploaded into an autonomous handled sized car. The RL model is then used in conjunction with the onboard camera, gyroscope, and accelerometer sensors to guide it around a racing track as quickly as possible. If your time is the fastest, you'll be at the top of the leader board. 

Anyone can join this league and you can purchase your own AWS DeepRacer to autonomous car, which is being sold through The car is a 1:18th scale vehicle which has a camera at the front as its main guiding sensor. The AWS DeepRacer autonomous car comes with the following specifications. The racing league itself is comprised of competitions hosted at each local AWS Summit. Winners of these competitions graduate through to the grand final to be hosted at the next AWS ReInvent conference. Additionally there are virtual events where you can upload your RL models and race within a simulated environment. The virtual events will involve a simulated racing track that progresses in complexity over time. 

When building and training your Reinforcement Models, you'll use the AWS DeepRacer console. Here you'll establish a training job which uses a supported RL framework, algorithm, reward function, and other hyperparameters. The reward function is central to the entire outcome and is something you must develop and invest time in correcting and optimizing. The reward function will ultimately determine how fast your autonomous car can navigate the course correctly, and the reward function is implemented as a Python based script, and the logic needs to consider many input variables, a few examples being: the X and Y car location coordinates, on or off the track, displacement from the center line, the car orientation, the percentage of track completed, the number of steps completed, the speed of the car, and the steering position. The AWS DEepRacer console provides an example reward function that can be used as a starting point. It can be edited and customized, or completely replaced with your own advanced custom logic. Having completed the design of your reward function, and with it uploaded you're ready to train your RL model. 

Once the training has completed, you can evaluate the performance using the simulator within the console. The evaluation itself involves racing X number of laps. A measure of performance is determined by calculating the average time to complete each lap. Once your RL model has achieved a desired performance level, download the model artifacts locally and next connect the autonomous DeepRacer car to your workstation using a USB cable and then upload the model artifacts to the car and you are ready to burn rubber. Getting involved with AWS DeepRacer will help you learn and gain valuable experience with reinforcement learning, while having fun at the same time. Your attention span will now be measured in days if not weeks. 

That now brings me to the end of this course covering Amazon SageMaker Ground Truth, Amazon Forecast, Amazon Comprehend Medical, Amazon Textract, Amazon Personalize, Amazon SageMaker RL and Amazon DeepRacer from the machine learning category. If you have any feedback on this Re:invent Reminder course, positive or negative, please contact us by sending an email to Your feedback is greatly appreciated. Thank you for your time and good luck with your continued learning at Cloud Computing. Thank you.

About the Author
Learning Paths

Stuart has been working within the IT industry for two decades covering a huge range of topic areas and technologies, from data center and network infrastructure design, to cloud architecture and implementation.

To date, Stuart has created 150+ courses relating to Cloud reaching over 180,000 students, mostly within the AWS category and with a heavy focus on security and compliance.

Stuart is a member of the AWS Community Builders Program for his contributions towards AWS.

He is AWS certified and accredited in addition to being a published author covering topics across the AWS landscape.

In January 2016 Stuart was awarded ‘Expert of the Year Award 2015’ from Experts Exchange for his knowledge share within cloud services to the community.

Stuart enjoys writing about cloud technologies and you will find many of his articles within our blog pages.