Quick Take

We helped Mylo as its founding technology partner, launch the new concept through an MVP responsive web application, backed by Amazon SageMaker. Mylo is harnessing the power of machine learning to mobilise people through learning their spending habits and getting thousands of millennials savings for their future.

What We Did

  • From concept to marketplace.
  • Containerised cloud-native application architecture.
  • Machine learning with Amazon SageMaker.
  • Modern development with Angular, Node.js, and PostgreSQL.

The Challenge

Most financial advisors - people who specialise in identifying opportunities to increase your net wealth by investing savings into complex financial products, are extremely expensive and not easily accessible by everyone. The founders of Mylo which stand for Managing Your Left Overs; approached OSO DevOps for all-around technical guidance on how to build a minimum viable product using the latest technology as cost-effective as possible.

The brief was to gain a deep understanding of customers spending, but highly valuable data sources, such as transactional card data, were not accessible through existing transitional APIs. This meant that Mylo could not easily consume the data it needed to learn customer spending habits and provide saving recommendations.

We knew we wanted to create a reusable, scalable machine learning model on AWS. So we reached out to our partner manager at Amazon to see if they’d be interested in working with us. They agreed enthusiastically and helped fund part of the project in the form of platform credits.

The Solution

As a Select AWS Consulting Partner, OSO DevOps was selected to complete the design and development of the product backlog. AWS offered agility and depth of services unmatched by other cloud-based providers, the environment is PCI compliant, using best practices and built to scale and grow as adoption increases.

The platform was built using a variety of services including Airflow for ETL, real-time feed integration with Yodlee, S3 buckets for organised data storage, SageMaker for analysis, as well as CI/CD tooling for automated deployments. The solution provides for self-service business insights and reporting functionality to other business units such as Customer Service.

Cloud services & technologies leveraged:

  • GitHub – SCM used as the code repository with commit hooks to trigger builds.
  • Hashicorp Terraform - Used to automate the infrastructure including network, security and application definitions.
  • Apache Airflow - Used for creating ETL workflow's.
  • Amazon ECS Fargate – Application infrastructure scaling by utilising Application Load Balancing to the ECS Fargate service.
  • Amazon CodeBuild & CodePipeline - Created a Docker image build pipeline to harden the underlying image and install applications.
  • Amazon S3 – Store static content including images, stylesheets, artefacts, and configuration.
  • Amazon IAM – Access control of users across accounts and environments; Use of IAM Roles to limit the use of access keys in implementations.
  • Amazon Lambda – Used for automating tasks and linking services together.
  • Amazon Athena – Used for querying the data stored in S3.
  • Amazon SageMaker – Used for creating data analysis models.

The Business Outcome

As a direct result of OSO DevOps engagement, it has enabled the FinTech startup to offer personalised saving recommendations, unique spending insights for the customer and exponentially higher amounts of money saved. Due to the success of the first stage of this project, additional funding has now been secured for future product development and scaling the business. Finally, by baking security and governance guardrails into the software delivery pipeline, compliance was guaranteed from the outset and risks were reduced, while maintaining high throughput.

Speak to an expert today

Contact us today and one of our experts will get in touch to talk you about our services and solutions or answer any questions you may have.