How do Headspace & Exo utilise Machine Learning to reshape healthcare?

How do Headspace & Exo utilise Machine Learning to reshape healthcare?

Natalia Figas
August 31, 2021  | 4 min read

Faster and better-tailored health services? Headspace proves that this can be achieved with Real-Time Machine Learning, leaving the competition far behind.

Experts claim that the pandemic of COVID-19 accelerated the growth of healthcare technology trends such as telemedicine, 3D printing or Internet of Medical Things (IoMT). Adaptation of new business models followed during almost one year. Digitisation is a step that has been delayed for far too long, however companies in this sector are catching up quickly.

An example is the Digital therapeutics (DTx) category and the Headspace app, which is leading in this field.

Will Digital therapeutics (DTx) become a new standard for healthcare services?

Whether we think of an end-of-life care, chronic diseases or treatable diseases, we have an ongoing healthcare on our minds. And we expect the highest quality therapy.

Digital therapeutics, also known as “Software as a Medical Device” (SaMD), are evidence-based therapeutic interventions driven by high-quality software programs to prevent, manage, or treat a medical disorder or disease.

These programs acquire data through mobile and web applications, software systems and Internet of Things (IoT) devices. Then they utilise Machine Learning to support a treatment in real-time.

On the contrary to a non-digital therapy, all activities and results can be monitored and measured right away.

To ensure the highest standards of efficacy and safety, DTx products undergo rigorous clinical and regulatory reviews.

What can you treat with Digital therapeutics?

So, what does DTx helps to treat? A broad spectrum of medical disorders or diseases. Below you will find a few examples. In brackets, we put DTx apps known for their positive impact on patients’ health. All of them are highlighted by ORCHA, an agency analysing the digital therapeutics market.

However, remember, the market of the Digital therapeutics is young and grows rapidly. Surely, in a few years it will become a new standard in healthcare services. But in order to ensure it, one main factor must be met: a digital transformation of healthcare.

How is DTx reshaping healthcare?

Patients’ needs and possibilities go through an ongoing change. Healthcare managers need to develop novel care delivery models in order to follow these changes.

Physicians can utilise the Digital therapeutics in two ways:

  • as stand-alone treatments or
  • alongside existing drugs or therapies.
Artificial intelligence (AI) involves the use of adaptive algorithms to perform tasks which typically require human cognition. If harnessed thoughtfully, both DTx and AI can make palliative and hospice care more effective, efficient, and expansive
- Journal Of Pain and Symptom Management

As we said before, technology which we use every day, such as mobile apps, wearable devices, web platforms – all of these empower the DTx. To illustrate how Machine Learning reshapes healthcare, let us take a look at Headspace where Real-time Machine Learning is being utilised.

Real-Time Machine Learning at Headspace

Headspace’s users improve their health through mindfulness, meditation, sleep, exercise, and focus content. This mobile and web application stands out of the crowd by offering personalised content that builds consistent habits. Moreover, Headspace does this within seconds!

Machine learning models at Headspace

Traditionally, a training of ML model involves a few steps:

  1. Data collection
  2. Data preparation
  3. Choosing and training a model
  4. Evaluating the model
  5. Parameter tuning
  6. Making predictions

Some of these are executed via periodic jobs that run every few hours or daily. But Headspace upgraded a traditional path.

Their Machine Learning team significantly shortened the time to deliver deeply personalised content. The user receives recommendations in real-time - right away when they are still using the app.

How did Headspace do this? They leveraged some specific solutions (Apache Spark Structured Streaming on Databricks, AWS SQS, Lambda, and Sagemaker) and combined them into a specific architecture (decomposed the infrastructure systems into modular Publishing, Receiver, Orchestration, and Serving layers) to deliver real-time inference capabilities for their machine learning models.

Another great case study of using Machine Learning in healthcare comes with Exo Imaging.

Exo’s revolution

Our client, Exo Imaging, utilised Machine Learning and AI to build an ultrasound platform for medical imaging. Thanks to Machine Learning and appropriate data collection the company co-created a coaching program for young cardiologists. helped transform the knowledge and experience of senior specialists into digital training.

Collecting data - the most important fundament for Machine Learning

An echocardiogram can help detect damage caused by a heart attack, heart failure, congenital heart disease, problems with the heart valves, cardiomyopathies and more. But to discover abnormalities a doctor needs to know what to look for.

To create software that would support young medics in their learning path, Exo needed a proper probe for data collection and a vast amount of data from heart examinations.

In this project, were to:

  • find medics, hospitals and universities eager to deliver examination data,
  • get permissions to use the data,
  • collect a huge amount of data,
  • process and label it properly,
  • and create an intelligent system with an intuitive UI that would teach the algorithms about heart examinations,
  • test in real-life if our product would work.

In short, the software’s algorithms learn from data that we collected with experienced operators.

In order to provide modern end-user applications, we refactored the existing system. Moreover, we extended it with components and developed new features. But that is a separate story, which you can read here.

R&D and data labelling

Our R&D experts examined the possibilities of sensors of the cardiographic probe. They learned how the device collects information during the test, at what angle sensors are set, how quickly the doctor works with the probe, etc. With such a knowledge, we were prepared to start the next step: data labelling.

We prepared an advanced data labelling system that immediately recognised an image category and content, labelled it and created a base for further learning of AI.

To work, the system needed integrations with medical devices for data downloading. The Skyrise’s team also prepared and tested those integrations.

Working with data

Working with data is at the core of machine learning algorithms. A sufficiently large and selected group of data is needed. And that data should be properly labelled.

If you are looking for a tech partner who can do it all, let's talk about your project.


At, we create custom digital solutions that meet your needs. We work for enterprises or industries such as logistics, but we focus on the healthcare market.

We are a part of Etteplan, a Finnish engineering company. Together with our colleagues from Denmark, Sweden, the USA, China, the Netherlands and Finland, we are a team of 3,400 specialists ready to add value to your project.