The tech that allows collaboration without compromise

© 2020 Manuel Geissinger

© 2020 Manuel Geissinger

Data is all around us. Companies, governments, and healthcare all rely on data to make smart decisions. But what if that data is incomplete or wrong? Tune Insight, a startup created by Dr Juan Troncoso-Pastoriza and Professor Jean-Pierre Hubaux, helps businesses make better decisions thanks to secure data collaboration. Dr Troncoso-Pastoriza discusses the potentially limitless applications of this Innogrant-funded technology.

Data is big business and it’s growing. Every day, the world creates 2.5 quintillion bytes of data. To put that into context, every minute of every day we make over 3.5 million Google searches. We watch over 4 million YouTube videos. We send over 16 million text messages. 

Nearly everything we do - from catching up on our favourite Netflix show, to tracking a walk on our Fitbit - creates data. In fact, 90% of the world’s data was created in just the last two years. By 2030, it’s estimated the data economy will be worth $5 trillion.

Organizations are growing increasingly dependent on data – using it to make strategic decisions about their business. However, there’s a problem. Despite the huge amount of data out there, due to security concerns and data laws and regulations, companies can only access their own information to make decisions. Insufficient and incomplete data means risky and potentially wrong decisions: Their own data is not enough.

What if there was a way for companies to securely collaborate around their most sensitive data, gaining valuable collective insight while remaining in control of their own data?

Tune Insight, a new venture incubating at EPFL, has the answer. The startup, founded by Dr Juan Troncoso-Pastoriza and Professor Jean-Pierre Hubaux, provides software which enables organisations to securely share data, unlocking the insight and intelligence they need to grow their business. We spoke to Dr Troncoso-Pastoriza to find out more…

Tell us a little bit of the Tune Insight story

Our dream is to help companies collaborate around their most sensitive data in a way where everyone wins – the companies, their customers, and the providers of the data.

Every company or organisation collects data as part of their daily work. Depending on the industry, this may be data about their customers, information about competitors, or figures on the impact of their products (e.g., drug efficacy). But there are limits on the amount of data one company can collect, both from a regulatory standpoint and from a practical standpoint. This means that they may have a limited view of the world around them and are unable to use precise insight to plan future strategy.

With Tune Insight, we have an opportunity to help companies broaden that view, giving them access to a high-quality and accurate insight based on data from large-scale networks, enabling them to make better business decisions. 

The main barrier to data sharing is confidentiality and control. No one wants their data to be shared without the proper protection and guarantees in place. Some of the data we’re talking about is very sensitive, very confidential, data which would cause real harm if it were leaked. Some of the data may be directly connected to the competitive advantage of a company. We’ve developed a software that relies on sophisticated and efficient encryption protocols and means data can be anonymised and protected, without losing its intelligence value. 

What does this mean in practice? How might Tune Insight software be used?

One of the best examples – and something that’s particularly relevant in the era of COVID-19 – is in clinical and health research. 

Researchers are currently trying to gather as much information as they can about COVID – who it affects, how it affects them, and potential vaccines and cures. A true picture of the behavior of the disease and its impact on the population can only be achieved if these researchers have access to data from patients all around the world. Data protection regulations, security risks and other legal constraints mean it’s currently not possible to quickly and effectively share patient data between hospitals and, especially when it’s across borders. The current mechanisms for ethics approval take into account many factors, including patient consent, and how to guarantee the respect for consent and for data protection. This process can be lengthy. It’s hampered by the fact there are no common tools available to guarantee this data protection. These regulatory hurdles can slow down important research. While some data can be sufficiently anonymised with traditional (and mostly manual) mechanisms, data used for medical research, like genomics, can’t. It’s not possible to separate the identity from the data without rendering it useless. The only way of sharing this important data is by relying on cryptographic techniques like the ones we use at Tune Insight. Our technology enables companies to protect sensitive patient data – including genetic information – so it can be used for crucial medial research and diagnosis. 

Imagine a patient presents with certain symptoms at a hospital. Usually, doctors only have a small dataset of other patients from that particular hospital to use to help diagnosis. With Tune Insight, they would be able to get much broader insight from data in a large network of hospitals – aggregated information on how patients with similar features or similar genes reacted to different treatments or drugs. With this information, the doctor could evaluate the best performing treatment for that specific type of patient. 

Is that the next step? Personalised medicine?

Every medical professional wants to have the right information available to be able to make better treatment decisions and deliver more personalised care. One of our tools, MedCo, has already been tested and deployed in Swiss university hospitals. It makes up part of the Swiss Personalized Health Network helping these hospitals with their clinical research on oncology and infectious diseases. Our software has a number of applications beyond medicine though. We are actively working and testing the use of our technology to enhance insights in cybersecurity and cyberdefense networks in collaboration with armasuisse. Likewise, insurance companies could use aggregated data to deliver more accurate premiums, banks could use more intelligence to make more accurate market predictions, and telecom operators could use our system to better manage their networks. The possibilities are endless! 

How will the Innogrant help you?

I’ve worked in academia for more than 15 years now, and the industrial world is very different. The Innogrant will allow me to transition from academic tasks, like teaching and writing papers, to focus on what we can achieve with Tune Insight. The Innogrant will allow me to take the time to be in touch with the adopters of our solutions. I want to mature and advance our technology so it can respond to the real demands and make an impact - coming up with real solutions to real problems. What’s the end goal? We’re aiming at the moon! This technology isn’t restricted by geography; on the contrary, it is an enabler for cross-border collaborations. The solutions we’ve already tested and deployed in Switzerland can be applied almost anywhere else in the world. We’ve already started a project to enable privacy-conscious COVID-19 research across the globe – Korea, France, Nigeria, Italy, Switzerland and the US are already participating. 

Everyone in our lab has a pivotal role in the past and future success of our technology. It’s thanks to this combination of really great minds and development skills that we have this opportunity to make our ideas a reality and make a real and substantial difference to people and organizations all around the world.


Tune Insight is a start-up in incubation based in EPFL’s Laboratory of for Data Security led by Professor Jean-Pierre Hubaux. Find out more information on their website

For more information on Innogrants and details on how to apply, visit the Startup Unit website.