Product
Product
Health Library
Health Library
Calculators
Calculators
About
About
March 09, 2021

Harnessing Machine Learning and Amazon SageMaker to Advance Women’s Health

Founded in 2015, Flo aims to improve the health and well-being of every girl and woman worldwide. Because the body can present as many mysteries as miracles, Flo was created to streamline female health expertise in one easy-to-access location. 

Now, people everywhere can use Flo to get expert, evidence-based, and personalized information about their unique health queries safely and securely. To actualize this ambitious mission, we’re scaling quickly to match the demands of over 165 million users while maintaining an innovative ethos. That’s why we chose Amazon Web Services (AWS) — one of the most globally trusted and audited cloud computing providers available.

Female health has received more attention in recent years than at any point in history. However, there’s still much work to be done. Female-centered health brands continue to face significant barriers, including variation among clients. Everyone is unique — no two people have identical cycle patterns or body reactions. It’s very hard to build a one-size-fits-all solution for a population from different geographies and with different health conditions. To meet the needs of this diverse community, Flo intends to expand its reach and services through machine learning (ML) and Amazon SageMaker capabilities. 

We’ve backed our efforts with a few core promises: accuracy, personalization, and security. This means providing accurate health information — including cycle predictions — delivering personalized content, and keeping user data private and secure. To better fulfill these promises, Flo is taking great technological leaps.

Using machine learning to provide accurate cycle predictions and highly personalized content 

Instead of simply combining data from a user’s most recent cycles, Flo uses ML techniques to make accurate cycle length predictions based on the information previously logged by the user, including their age, period dates, symptoms, and other insightful metadata. Flo’s latest milestone includes the development of a new MVP for a new feature that can potentially foreshadow symptoms. The Flo team is also investigating other ways to harness ML to predict ovulation and fertile windows with ever greater accuracy and precision.

In addition to these cutting-edge developments, Flo continues to provide a plethora of personalized multimedia content that has helped millions of users to better educate and understand themselves. The Flo content library includes video courses, articles, stories, and more. This content is reviewed by over 80 renowned health and well-being experts to ensure that all of the information is evidence-based and trustworthy, and smart algorithms help pair users with personalized, relevant information. In addition to these resources, users also have access to an anonymous community, called Secret Chats, where they can connect with each other about shared experiences, questions, concerns, and support. This allows users to address their unique personal needs from every angle. 

Scaling safely and quickly with Amazon SageMaker

Although the Flo app was originally built on a smaller server, we migrated to a cloud provider after realizing that more elasticity and a solid infrastructure were essential to scaling quickly. We decided to go with AWS because it’s a reputable and trusted provider that would help us leverage one of our overarching values — building a safe, dependable, and secure space for millions of people worldwide. “From the moment the company was founded, we have been consistently working to create an absolutely secure product for our users,” says CTO Roman Bugaev. “To that end, AWS gives us and other companies a secure space to store the encrypted data safely.”  With a growing number of engineers working with data from Flo’s over 165 million users, AWS provides a platform that allows the Flo team to focus on innovation.

“The data science department specifically relies on Amazon SageMaker for ML, leveraging powerful GPU-based machines offered by AWS to train models on Flo’s large user population,” says Bugaev. He also emphasizes that working with AWS is bidirectional. The Flo team uses technology to the greatest extent of its capabilities, and the AWS team responds to our feedback in real time. “It’s great to have a trusted provider who can support us with best-in-class services, so we don’t need to reinvent them ourselves,” says Bugaev. “It allows us to focus on our core competency, which is improving the health and well-being of every girl and woman worldwide.

Source: AWS Startup Blog