In September, I had the opportunity to participate in the AWS BlackBelt Program, in the stack of Machine Learning and Artificial Intelligence. In this post, I will share with you a little about my experience there. But before describing the activities, I will first explain the program.
The AWS BlackBelt is a program created by AWS in South America, and it has the purpose to create, train and develop a group of specialists in AWS in different stacks (Machine Learning, Analytics, Database, Enterprise Migration and Security), so these people can use the latest AWS technologies when doing projects for external clients.
The participants are selected from partner companies and there are some requirements that must be met by the candidates (such as AWS certifications) to prove they do know about the stacks they are participating in. Each year, there are two editions – this year was the first one – and in each edition, the requirements for new members (and for current ones) increase, so that the group becomes increasingly specialized.
The program lasted 3 days. The first day started with some discussions about what had changed in Machine Learning on AWS, focusing especially on SageMaker (you can read more about it here). After, there was a discussion about how to sell Machine Learning projects – and its difference from “pure” IT projects –, moment when all members of the group shared their experiences. The afternoon started with a demonstration of SageMaker NEO, a tool to optimize trained Machine Learning models, in order to use less power/memory to make inferences and deploy them to devices with less computational power, something especially important for IoT. Finally, the day ended with discussions about recent advances in the field of Machine Learning.
The goal of the second day was to get our hands dirty solving some problems. The group was divided into pairs and each pair had to use Machine Learning to solve a problem using the AWS structure. The problems were varied, ranging from image classification and segmentation, sales forecast, recommender systems, to deploying a face recognition model on a Raspberry PI. At the end of the day, each pair presented its work for the group, discussing what had been done and why.
In the first two days, I stayed only with the members of the Machine Learning BlackBelt. The dynamic of the third day was different though: all the participants were reunited, and mixed groups of 6 people were formed, with each group having a specialist of each stack. Then, all the groups participated in a game day, where we had several challenges to solve in a specific timeframe. The challenges were varied, ranging from deploying a website in S3, creating a face recognition application by using AWS Rekognition and Lambda, to querying some data files using Athena.
This experience was very helpful for me, because it allowed me to connect with new people and to see the several different ways that one can structure a Machine Learning project. It also helped me to see how to use the different AWS tools to deliver maximum value for clients in terms of Machine Learning. These two things combined will certainly help me and Poatek to deliver even better projects in the future.
About the author
Luiz Nonenmacher is a Data Scientist at Poatek. He has a Master’s Degree in Production Engineering in the area of Machine Learning and Quantitative Methods. On his spare time, he likes to read about a lot of different subjects, including (but not limited to) classical philosophy (especially Stoicism), science, history, fantasy, and science fiction.