In this workshop, you will learn how to use Amazon SageMaker to build, train and host machine learning models. Going through a number of Jupyter notebooks, you will first learn how to use built-in algorithms to perform complex tasks like image classification or clustering. Then, trainers will teach you how you can bring your own Tensorflow or Apache MXNet script to train deep learning models. Finally, you will deploy your models to SageMaker-managed infrastructure and use them to predict new samples.

Topics Covered:

• What Amazon SageMaker is
• How to build end-to-end machine learning workflows on Amazon SageMaker
• How to use built-in ML algorithms for classification, image recognition, etc.
• How to bring your own model, your own training code, etc.

And more..

Who should attend?

Technology leaders from Startups and Enterprises (Engineering, Architecture, Product, Development) who are interested in expanding your knowledge on Artificial Intelligence, Machine Learning, and how it can be applied to your business.

Note: This workshop is sponsored by AWS and the data collected for this workshop will be shared with them.

Agenda

1. Overview of AI/ML/DL [30 mins]
2. Amazon SageMaker overview – an overview of the SageMaker service, best use cases, main features including AWS security concepts of IAM, VPC, KMS. [ 45 mins.]
3. Accessing SageMaker – demo to show how to easily access SageMaker service [Duration: 15 mins.]
4. Notebook demo on using highly -optimized built-in Amazon algorithms [Duration: 30 mins.]
5. Hands-on lab – Managed Training, Hosting and A/B Testing of Amazon built-in algorithm – Amazon linear learner algorithm / parallel training using SageMaker Estimators / SageMaker Python SDK [Duration: 45 mins.]
6. Hands-on lab – Build your own DNNs using MXNet/Tensorflow, distributed training on GPUs and serving using SageMaker [Duration: 1 hr. 15 mins.]
7. Integration with Amazon Elastic Map Reduce (Managed Hadoop Service) - Amazon SageMaker notebooks backed by Spark in Amazon EMR [Duration: 1 hr.]

Pre-requisites

• Participants should carry own laptop.
• To participate in hands-on sessions, you need to have an AWS account. If you don’t have, please create one. You will be required to share credit card details to validate your identity. Please do create an account now as it take sometimes few hours to validate. We will be giving every participant 100$ credit for this workshop. At the end of the workshop, you should terminate all Sagemaker resources created for the workshop (notebook instance, inference endpoints etc.) and delete all workshop related data from S3 to avoid unnecessary AWS Billing.
• There is a hands-on session in which you will be doing DNN training with GPUs. To follow this session, participants should have an AWS Account with admin privileges in IAM and EC2 limit for p2.xlarge instances increased to 2 in AWS Region North Virginia (us-east-1). You can mention “to train Resnet model with Tensorflow on 2 p2.xlarge instances” as usecase in the form. Check out this doc to know more about how to increase EC2 limits. All participants will be provided AWS Credits for the workshop. It usually takes 24 hours to validate. So please do this asap.
• Participants should be familiar with AI / ML / DL and need to be hands-on practitioners.