Personalized medicine software on AWS: 60% decrease in post-release bugs

Innowise engineered a serverless infrastructure on AWS, enabling health recommendations for personalized medicine software and implementing a robust CI/CD pipeline for seamless deployment and testing.

Customer

Industry

Healthcare IT

Region

Israel

Client since

2022

Our client is an innovator in the healthcare technology space. They focus on helping individuals with chronic conditions to achieve better health through following tailored AI-driven recommendations. With an emphasis on individualization, the client offers personalized medicine software and a mHealth app for individuals and healthcare providers.

Detailed information about the client cannot be disclosed under the provisions of the NDA.

Challenge

Redesigning the personalized medicine software to build a robust infrastructure for health recommendations

Our collaboration with the client, initiated in 2022, centered around enhancing their health management system available on web and mobile platforms. Utilizing AI and emotional intelligence (EI) technology, the personalized medicine software dynamically tailors itself to each user’s unique personality, habits, and lifestyle. This approach aims to facilitate sustainable lifelong changes and encourage adherence to treatment programs.The client’s challenge was redesigning a system to track user actions and deliver health-optimizing recommendations more precisely. It required a robust infrastructure for handling recommendations and notifications, which had to be scalable and maintainable. Furthermore, both web and mobile platform versions were lacking comprehensive QA services.

Solution

Personalized medicine software with improved back-end system, A/B testing, and streamlined CI/CD pipeline

In our engagement with the client, we developed an advanced and scalable infrastructure to support personalized medicine software with health recommendations. “Emphasizing Infrastructure as Code (IaC) practices, we combined the power of AWS CDK with TypeScript. This enabled us to create a robust, serverless framework capable of handling complex recommendation and notification processes, integral to enhancing health management. Our team also focused extensively on testing mobile applications and back-end systems.

Infrastructure deployment

In the foundational phase of our development process, we deployed the AWS Cloud Development Kit (CDK) with TypeScript. This strategic choice allowed us to script cloud infrastructure as if it were software. It streamlined the creation of resources and ensured that our setup was maintainable and easily scalable — crucial for a system designed to handle a growing number of users.The infrastructure’s backbone was a serverless architecture designed for high availability and cost-efficiency. Serverless computing enabled us to build and run applications without thinking about servers. This meant we could focus on the core product without the overhead of managing infrastructure.To weave together the diverse actions and services the application requires, we leveraged AWS Step Functions to create State Machines. They ensured that each user interaction was processed accurately, triggering the correct sequence of events — from data ingestion to personalized JIT (Just In Time)-notifications and recommendations.

mHealth app development

Our team embraced Python for its simplicity and efficiency in crafting Lambda functions, which formed the core of various back-end services. These functions were responsible for the entire user notification process, from data processing to sending out health recommendations and messages.Our PHP development team enhanced the UI of a web application using October CMS, focusing on a user-friendly system for managing notification data, including templates, placeholders, and conditions. This shift from a developer-reliant model to a more accessible approach has empowered product owners to manage data independently, streamlining the workflow and increasing efficiency in the notification management process.The complexity of transforming raw user data into insights called for robust ETL processes. AWS Glue provided a managed ETL service that simplified the preparation and loading of data for analysis. To manage the data flow in real-time, we constructed Data Pipelines incorporating services like AWS EventBridge for event bus routing and AWS Kinesis for handling massive health data streams, ensuring that user interactions were processed and acted upon without delay.

Continuous integration and deployment

To maintain consistency and quality in our deployment processes, we established CI/CD pipelines utilizing Bitbucket for source control and AWS CodePipeline for orchestrating the builds, tests, and deployments. These pipelines facilitated a smooth transition from development to production, with automated steps that reduced human error and streamlined releases.

Amazon Elastic Container Service (ECS) was configured to run and manage our Docker containers. This service simplified the system’s container orchestration, allowing us to deploy, manage, and scale the recommendation and notification systems with ease.

Quality assurance

Quality assurance was a critical and integral component throughout our deployment process. Our QA engineers validated the back-end system’s functionality, performance, and usability and ensured the mobile application’s highest quality by combining manual and automated testing methods.

A key focus of our QA strategy was the thorough testing of the personalized medicine software on mobile platforms. We conducted extensive manual testing by simulating real-world user scenarios to ensure the app’s interface and features worked flawlessly on various devices. This was complemented by running automated tests to cover a broader range of use cases.

Managing the CI/CD workflows was another vital aspect of our QA process. We monitored these workflows to prevent untested or buggy code from being deployed to production. This approach became especially crucial after identifying process gaps that allowed bugs to emerge in the live application, particularly during the critical 2.0 release for the new market.

To further refine the application based on user interaction, we implemented A/B testing mechanisms. This enhanced user engagement and provided valuable insights into user behavior and preferences, enabling the client to make data-driven improvements to their product.

The client was particularly impressed with the robustness of our mobile and back-end testing, as well as the efficiency of the CI/CD pipeline. These efforts led to a significant decrease in deployment-related issues and a substantial increase in the mHealth application’s stability.

Technologies & tools

Back-end

PHP, Python, TypeScript

Cloud

AWS (Step Functions, Lambda, Kinesis, Event Bridge, Api Gateway, CloudFormation, Glue, Athena, App Sync, ECS, ECR, Batch, RDS, Redshift, DynamoDB)

Databases

Postgres, Redshift, Redis, DynamoDB

Source controls systems

Bitbucket

Pipelines

Bitbucket Pipelines, Code Pipeline

Process

Our work with the client was marked by step-by-step progression, transparent communication, and a strong commitment to Agile methodologies. This approach enabled us to adapt quickly, maintain consistent engagement with the client, and continuously improve our processes throughout the project. Here’s how the project unfolded:

Initiation and planning

We commenced with a thorough analysis and planning phase, aligning our tasks with the client's needs. This stage set the groundwork for what would become a responsive mHealth app development cycle.

Infrastructure setup

Using AWS CDK, we scripted the infrastructure to support a serverless back end, ensuring the system was scalable and resilient.

Development of functions

Our developers wrote Lambda functions to process data and handle notifications, managed through the serverless infrastructure.

CI/CD pipeline construction

We set up Bitbucket and AWS CodePipeline to automate the deployment process for infrastructure and applications.

Quality assurance

Our QA engineers conducted thorough manual and automated tests to ensure all features worked properly across different devices and user scenarios.

A/B testing implementation

To further enhance the user experience, we established an A/B testing framework, allowing for data-driven decision-making.

Project completion and review

The project concluded with a comprehensive review and handover phase. We ensured that all project elements met the client's expectations and prepared the groundwork for future enhancements and support.

Team

AWS Developers

PHP Developer

Python Developer

DevOps Engineer

QA Engineer

Project Coordinator

Results

Enhanced backend system, 20% reduction in time-to-market for new features, and enhanced stability for the personalized medicine software

The collaboration with the client led to several notable achievements, each contributing to the overall success and impact of the personalized medicine software:

Enhanced back-end system: We developed a robust, scalable, and cost-efficient back-end system. The serverless architecture on AWS allowed the application to handle increasing loads and user data efficiently.
Improved application stability: We significantly reduced critical bugs through rigorous QA processes. The application’s stability was enhanced by approximately 40%, as indicated by the decrease in crash reports and user-reported issues.
A/B testing for continuous improvement: Implementing an A/B testing framework marked a key milestone. It allowed the client to fine-tune the application based on user feedback and behavior, improving user satisfaction scores.
Operational efficiency: Automating deployment processes and introducing a more streamlined CI/CD pipeline reduced the time-to-market for new features by 20%.

In summary, our QA-oriented approach and the robust AWS serverless architecture provided our client with a highly reliable and personalized medicine software. These improvements supported their mission of delivering personalized healthcare solutions, as evidenced by the tangible enhancements in application performance and user satisfaction.

In the current phase of our project, our dedicated team is actively engaged in ongoing mHealth app development and enhancement, with a strong emphasis on testing and continuous improvement of the infrastructure.

Project duration

January 2022 - Ongoing

20%

reduction in time-to-market for new features

60%

decrease in post-release bugs

Related cases

Contact us!

Book a call or fill out the form below and we’ll get back to you once we’ve processed your request.

What happens next?

Having received and processed your request, we will get back to you shortly to detail your project needs and sign an NDA to ensure the confidentiality of information.

After examining requirements, our analysts and developers devise a project proposal with the scope of works, team size, time, and cost estimates.

We arrange a meeting with you to discuss the offer and come to an agreement.

We sign a contract and start working on your project as quickly as possible.

Personalized medicine software on AWS: 60% decrease in post-release bugs

Customer

Challenge

Redesigning the personalized medicine software to build a robust infrastructure for health recommendations

Solution