ML and AI Solutions in AWS

Photo by Christian Wiediger on Unsplash

Cloud computing is one of the many fields I have been reading about during these uncertain days. The fact that industries across the globe are continuing to function despite the lack of a majority of the employees on-site is majorly due to the services provided by many cloud providers. Today, the three main cloud providers are Amazon (Amazon Web Services), Microsoft (Microsoft Azure), and Google (Google Cloud Platform).

Although AWS started as a provider Infrastructure as a Service (IaaS), it has since expanded to more than 175 products and services and now is the holder of a large part of the market share. Its Machine Learning and Artificial Intelligence solutions are of great significance, especially in a world that increasingly depends on data processing and analysis for building predictive models.

The 14 Artificial Intelligence services are pre-trained and can be easily applied for a wide range of different applications like advanced text analysis using Natural Language Processing (NLP), demand forecasts, chatbots, personalized recommendations, etc. Similarly, there are 4 main Machine Learning services under the umbrella of Amazon SageMaker. Most importantly, these services are optimized so that the users can focus on building business solutions without worrying about their experience in these fields. There are dozens of users for these services across various fields like Data Analytics, financial services, customer experience, etc. Let’s look at some of the customers who have benefitted from these services of AWS.


BuildFax is a wholly-owned subsidiary of Verisk, which was started in 2008. It is based in Asheville, N.C., and provides critical information and solutions to insurance and investment firms.

AWS enhanced the predictive models of BuildFax which were used to provide cost estimates for annual roof losses for insurance companies. Previously, the models built using R and Python took more than 6 months to develop, and they often did not provide significant differentiators. Using AWS Machine Learning solutions, they now have a streamlined process to build models and easy access to the predictions of the models through APIs. The models themselves have higher accuracy than their predecessors, as they are now built on a wider dataset from public sources and customers and use property-specific values instead of the original ZIP-code level estimates.


SailDrone specializes in the design and manufacture of sail drones, which are wind and solar-autonomous surface vehicles. Saildrones are not powered by any internal engine; they are dependant on wind power for propulsion. They are used for collecting data in real-time on the oceans, which can aid many marine studies and marine industries. Its headquarters is located in Alameda, CA, USA.

Saildrone collects global environmental data in real-time using wind-powered ocean drones. AWS assists them by providing Machine Learning solutions to study and predict the behavior of fish stocks and their predators. Recently, Saildrone completed the first successful autonomous circumnavigation of Antarctica. To achieve this, they utilized the ML services of AWS to reduce the risk of collisions with icebergs. Also, the compute and storage services of AWS were instrumental in collecting huge volumes of data in real-time over a few days.

Saint Louis University

Saint Louis University (SLU) is an example of how AWS can cater to both corporate and the academic world. Using the help of both Machine Learning and AWS Professional Services, the university has begun using a chatbot platform to facilitate and improve student and parent experience. Named AskSLU, it allows the students to ask a range of questions about SLU using Amazon Echo devices spread across the campus. Interestingly, the chatbot is designed in such a way that the students receive the same answer even if they ask the question through different mediums like the SLU website or even text messages. The chatbot also caters to requests of information by the parents.

Now, the university is looking forward to enhancing the capabilities of AskSLU and bring a more personalized experience for the users.


AutoDesk is a California-based company that creates software solutions for a variety of industries like architecture, construction, manufacturing, etc. They utilize technologies like 3D printing, robotics, generative design, AI, etc.

In 2017, AutoDesk moved its data science and machine learning practices from on-premises to AWS. But a significant development was the improvement of the Community Match using machine learning models for better customer experience. It was meant to facilitate communication between AutoDesk experts and their customers. In Community Match, AutoDesk was able to filter out customers based on their requirements and match them with members of the community who have specific expertise.

To bring in new members and customers to the Community, AutoDesk used an AWS serverless architecture to create a knowledge model to understand user expectations better. The knowledge model was built using Amazon SageMaker and hosted on Amazon Elastic Container Service (Amazon ECS), a fully managed container orchestration service, using a transfer learning technique to create embeddings of customer questions from the forums. The newer developments did not require any reconstruction of the existing software infrastructure at AutoDesk, which made the whole process easier to handle.

AutoDesk also bunched the models together with a set of business rules using AWS Lambda to pair a customer with a given requirement and the subject-matter expert with the requisite expertise. AWS Lambda allowed them to implement their system without worrying about the provisioning or management of the servers. Also, they used AWS Step Functions, which allowed them to sequence AWS Lambda functions and multiple AWS services. The complete solution allowed AutoDesk to introduce machine learning with real-time monitoring, without excess resource overheads and better resource scaling. Community Match now uses other services of Amazon like Amazon Simple Notification Service (Amazon SNS), a messaging service that is used to create topics or logical groups on different types of products and delivers update notifications. As mentioned in the AutoDesk Case Study of Amazon, the benefits of AWS in AutoDesk’s Community Match are

1. Created a solution prototype in 1 week

2. Matched an incoming inquiry with an expert 32% of the time

3. Saw a 31% click-through rate

4. Motivated 16% of low-engaged customers who received a recommendation to reply on the forum

5. Improved customer service

6. Relieved pressure on existing servers


These are just some of the variety of applications where Amazon Web Services are being implemented in the world. In the future, these applications and stories will only increase in number as we transition to a world largely run by data, data processing, and the insights gained from data analysis.

ECE Undergrad | ML, AI and Data Science Enthusiast | Avid Reader | Keen to explore different domains in Computer Science

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store