Empowering Aiosyn's AI-Powered Revolution for Precision and Progress in Healthcare

To support Aiosyn in their ambitions, we helped them to migrate to Amazon Web Services (AWS). David Tellez, CTO of Aiosyn, shares his firsthand account of their public cloud journey.

Lees meer

About Aiosyn

Aiosyn revolutionizes the field of pathology by empowering pathologists to enhance accuracy, efficiency, and effectiveness in their work. Through innovative software solutions that integrate deep-learning algorithms into existing pathology workflows, Aiosyn elevates clinical decision-making to new heights. Their AI-powered technology aims to offer new perspectives, ultimately leading to improved patient identification and stratification in drug development programs.

What was the initial situation before starting with CloudNation?

At Aiosyn, we develop, certify and integrate AI-powered solutions for clinical pathology. We were a cloud-native company from the very beginning. We recognized the flexibility of using cloud solutions instead of our own hardware very early. However, as soon as we started processing medical images and training AI models for our first products, we noticed that our system at that time presented limitations in important aspects like scalability, availability, traceability, security, etc.

What was the challenge Aiosyn was facing?

We needed to build a robust foundation of our computing infrastructure in the cloud, a starting point from which we could develop our products further while supporting the fast growth of Aiosyn. In particular, we identified three main goals:

Establishing infrastructure as code (IaC) as the standard approach to cloud within our organization, in particular using Terraform.
Following the best industry practices in security and compliance. An example of this was configuring a multi-account setup within AWS to isolate/compartmentalize sensitive cloud resources.
Maintaining flexibility and development speed. We did not want to lose the ability to develop prototypes and test hypotheses quickly, which is crucial in a startup like Aiosyn.

Why did you choose AWS? And why for CloudNation?

We chose AWS for three main reasons. First, our team already had some experience with this cloud provider, so we could start building infrastructure without delay. Second, we analyzed several cloud provider offerings at the time, and could not find substantial differences among them for our use case. Third, AWS supported us from the very beginning through their AWS Activate program for startups.

CloudNation was an easy choice for us. After the initial contact, they were enthusiastic to help us. They provided a tailor-made proposal to address the three goals that I mentioned before. Their consultants really understood not only the technical challenges but also Aiosyn’s unique challenges as a young startup building AI-powered medical devices.

Which phases did you go through from initial decision to migration?

The first step was to define the scope, way of working, and work packages of our project. CloudNation’s consultants provided invaluable advice in this early uncertain phase, and helped shape the requirements of our system. Second, together with CloudNation, we set up a landing zone in AWS following best practices such as multi-account, SSO, 2FA, etc. We built the system to automatically deploy cloud resources with IaC. Finally, we transitioned into a ticket-based way of working, where we would build infrastructure in biweekly sprints together with CloudNation’s help. Despite the technical challenges and complexity of the tasks, we really enjoyed our collaboration and still use the infrastructure on a daily basis to this day.

What are your plans for Aiosyn and your cloud usage?

We are extending our platform to support the integration and deployment of our production workloads. We are CE-certifying our first AI-powered medical device for mitosis detection in breast cancer. This requires us to improve our infrastructure to offer seamless integration with our customers, while optimizing performance and lowering costs. In addition, we will move into a multi cloud offer where we will be able to deploy our products not only in AWS but in other cloud providers such as Microsoft Azure. Our cloud journey has only just started!

Do you have advice for others starting or working with cloud?

Working with cloud technology can be a daunting task, but it definitely pays off in the long term if you are looking for flexibility and speed but do not want to compromise the quality of your infrastructure. We have a few tips:

Start early with infrastructure as code. The learning curve is steep but you will quickly be able to benefit from best practices and open-source code.
Build reliable infrastructure. Do not over-engineer cloud resources at the beginning, but try to build reliable and robust infrastructure even if it takes a bit longer. Infrastructure as code helps tremendously to achieve this goal.
Get expert advice early. Secure support from your cloud provider (e.g. AWS) and collaborate with an experienced partner like CloudNation. There are many important decisions to make that will impact the quality of your system in the long term.
Have fun! Building infrastructure in the cloud is one of the most fun activities for our engineering team.