Customers want AWS to do more to democratize cloud and machine learning

Profile picture for user Sooraj Shah By Sooraj Shah May 31, 2018
AWS customers say Amazon should do more to democratize its technology such as machine learning as well as helping to build cloud skills

People raise hands to vote in conference hall - Photo by Edwin Andrade on Unsplash
Since the first release of its S3 storage service and pay-for-use Elastic Compute Cloud (EC2) in 2006, Amazon Web Services has played a huge role in the democratization of computing. But the easy availability of cloud computing that AWS kickstarted 12 years ago has often raced ahead of the skills needed to make effective use of these online resources. Many organizations struggle to not only find qualified staff to fill open IT vacancies, but also ensure their existing IT teams stay up to date with emerging technologies.

Some customers have begun to argue that AWS can’t focus just on getting people to upskill, but must also find ways to make its technology more accessible to its users. This is particularly as the cloud skills shortage doesn’t cover one specific type of role in enterprises, but several. Skills are needed for cloud migration, cloud management, cloud architecture, and the use of emerging technologies such as machine learning and AI.

One approach from AWS is to get organizations using some of its easy-to-integrate tools first – such as its cloud-based contact centre AWS Connect and its deep learning chatbot service Lex. NHS Business Services Authority (BSA) has been using both. It’s a way of getting customers to standardize on their services. says cloud architect Chris Suter:

AWS are making it easier for businesses and start-ups to get on the journey with them – but at the same time trying to tie them into using their own code.

How to democratize data skills

Skills are in especially short supply when it comes to data management and analytics. Matthew Fryer, chief data science officer at, suggests that AWS has gone a long way in making cloud more accessible, but says that more investment needs to be made into making the use of data easier. He explains:

On the data stack there has been a lot of investment on the train and deploy end of the stack, but it’s important to remember without data you can’t have machine learning, and so more can be done to make it easy to capture and track data that is compliant with GDPR, and understanding the data quality and lineage – those are the challenges that I don’t think are solved.

He adds that, while AWS has made so much of the infrastructure simpler through its services vision –  including better integration with other technologies and the ability to go serverless – the data side is the one area that is letting the company down:

It’s almost as if Amazon has made compute and the operations easy and the underlying components are starting to come together, but we have to make data easy.

The machine learning engineer

One area where AWS has made huge strides is in machine learning technology. Dan Phelps, chief architect at Travelex says that the use of AWS SageMaker (a platform that enables developers and data scientists to build, train, and deploy machine learning models) and DeepLens (a wireless video camera and API that shows users how to use the latest AI tools and technology to develop computer vision applications) means organizations don’t need 20 data scientists with 10 years of experience to build out machine learning. He says:

They are giving you the tools to build the tech without being experts – there’s no need for huge people resource and that’s what I want to see more of from Amazon. We’d rather have a cloud provider that helps us to easily leverage these cloud concepts without having to build the capability and bring in a giant skill set.

Fryer agrees, explaining that the journey was initially difficult for and that it needed very deep competencies to use some of AWS’ technology.

But he says that data scientists – because of how rare they are – should only be deployed in the most technical use cases where they can bring the most value:

There are then hundreds of other use cases where we can benefit the customer if I can empower my wider data engineers and create a ‘machine learning engineering function’ as automated machine learning plays into the strengths and skills of engineers.

He maintains that data scientists will always be needed, but by empowering engineers with these new technologies, they could incorporate them into applications from day one rather than as an afterthought.

In addition, organizations would be able to better understand how to use the technologies at scale, rather than just focusing on a data science function. This would allow every part of the enterprise to drive that change, he believes:

This would create a machine-learning driven business.

Building up cloud skills

This switch from focusing only on data scientists towards incorporating engineers is something Fryer can see happening in the coming years. Considering the strides AWS has already taken in making its technology more accessible, Fryer suggests that the company will listen to its customers and get it right in the end. But he would like to see more progress made by improvements in the technology itself, rather than AWS’ various cloud skills schemes.

Chris Hayman, head of UK & Ireland public sector at AWS says that AWS is working on programmes to create access to cloud skills for both the public and private sector. At AWS Summit in London recently, he emphasized the company’s pledge to improve the skills of 100,000 people in Europe throughout 2018.

The cloud provider wants to shoulder the responsibility of building up skills through its various programmes such as Re:Start, which helps school children, young adults and ex-military personnel develop cloud skills, AWS Academy which helps universities and enterprises to train their staff, and the AWS Educate programme, which provides universities and secondary schools with AWS credits and curriculum resources to help teach their students about cloud.

Of course Amazon is thinking of its own future as well as that of the wider community. By getting people to use the company’s technology from a younger age, they’re more likely to use AWS for years to come. It’s a pattern that’s familiar to students and employees who became accustomed to Microsoft software decades ago and continue to use the company’s products today – and one that many other vendors have sought to emulate.