AWS re:Invent re:Cap Pt. 1


Last week AWS held its annual re:Invent conference. This year’s re:Invent was reminiscent of the first 2 re:Invent conferences with a focus on innovation. We’ve already established that AWS is the premier public cloud vendor for running Enterprise IT workloads. This year it is back to defining the future of computing, and to a larger degree technology consumption in general. As a technologist, this year’s re:Invent was an inspirational event that many of us will recall as a point of reference in our careers in years to come.


Dr. Werner Vogels took the stage on Thursday and delivered a direct message. The way we interact with technology is going to change and AWS is helping deliver the tools for us to change it. He also delivered a strong message about backend architectures. If we are to change the way we interact with technology, our backend architectures will need to be modern architectures that are redundant, resilient, and performant. We will all need to stop using the same old building blocks we’ve grown comfortable with and “re:Invent” new ones to build the technologies of the future.

This year was full of new service announcements and feature enhancements. The announcements started on Monday, but the game changing innovation was announced in AWS CEO Andy Jassy’s Keynote on Wednesday. In the coming sections and future installments, we consider the most significant of these announcements and what it means for technology and the IT industry.

Machine Learning

AWS focus in 2018 will be Machine Learning (ML) and Internet of Things (IoT). This installment will focus on the ML announcements:

  • Amazon Translate – Translate is a service that provides neural machine translation. This service meets the rigor for the future of localization challenges by going beyond rule-based translation algorithms to deliver a more accurate translation that can take into account dialects and nuances found in our everyday language. Translate will be able to deliver these capabilities through its use of ML and Deep Learning (DL).

  • Amazon Transcribe – Transcribe will quickly become a core building block service for the future of technology interaction as described above. With Transcribe developers can translate standard audio input into text. For example, we can now develop devices that can record a foreign language and write the file to S3, use Transcribe to convert that audio to text, then use Translate to convert to our natural language. This is just one obvious use case. A little imagination will deliver new and more interesting use cases in the coming weeks and months.

  • Amazon Comprehend – Comprehend is a Natural Language Processing (NLP) service. Comprehend can analyze text using ML to extract text for organization and categorization. This is another core building block service of the future. The use case I can’t wait to try is to combine Transcribe and Comprehend to take recordings of our Sales people’s conversations with customers, turn that audio into text, then use comprehend to put the data from those conversations into our capabilities: Migration, Architecture, Data and Analytics, DevOps or Machine Learning. I will even be able to use Translate to capture the details of the conversations of our Central and South American customers.

  • Amazon SageMaker – SageMaker is the service that is most likely to change everything. It lowers the bar of entry for development of ML models and allows the deployment of models at scale across a variety of environments (to a device like DeepLens for instance). SageMaker allows developers and data scientists to work with data, build algorithm experiments and visualize output. This service vastly increases the number of people who can develop and consume Machine Learning. It is also a fully managed service, so no one has to go build a bunch of infrastructure to get you started. A simple example solution can be created in as little as a one-hour session.

  • Amazon Rekognition Video - Amazon Rekognition, (note sans video) the service that was announced in November of 2016, provides capabilities to detect and recognize things in a picture. This year we have been helping customers provide interesting insights to their business through the detection of items in photos. Rekognition Video opens up the same capabilities to video, so now we can find out if Kevin Bacon really appeared on screen with all the people the internet claims he did by streaming our videos into Rekognition Video and looking for him and all those other high-profile faces.

  • AWS DeepLens – Amazon recognized that with the release of all these fascinating new services, specifically SageMaker and Rekognition Video, all us tech-heads, developers, inventors, and makers would need a device to hone our skills in on. Now enter DeepLens, a palm sized HD camera device with Greengrass (Lambda), the ML services and enough computing capabilities to do some activities locally. You can find a future use case for these technologies in a blog I wrote back in March here

AWS DeepLens was given to the very few who were in the Machine Learning Session on Wednesday. Relus Cloud’s own Gary Butler had a seat in that session, so we have a DeepLens and we are exploring its uses. Gary built a cool app with the rest of the participants in the session that trained the camera to recognize when it was being shown a hot dog. From this session we discovered it takes only an hour from introduction to working demo, impressive ramp time. Want a Demo? Contact us here.

While AWS’ competition thought they were making gains on it in the innovation department they should think again. This year AWS released a mountain of innovation that, without a doubt, put it way ahead of the competition. AWS is built on a culture of innovation, and that culture has done it again. Build On AWS, and build it with Relus Cloud as your partner of choice.