Hyfe is an AI-powered cough identification and tracking technology, co-founded by Lulian Circo, Joe Brew and Paul Rieger. Hyfe enables users to passively monitor their coughing habits, to better understand their overall wellbeing. Hyfe’s technology passively tracks cough sounds made close to a smartphone, before using AI to confirm that the sound was a cough.

Over time, users can see how frequently they cough through a dashboard on the Hyfe smartphone app. This can help users get insights into their base-level coughing habits so that they can spot anomalies as quickly as possible. The platform can help users identify symptoms worth raising with their doctors or help individuals understand the impact that air pollution is having on their health.

Tell us about the founder behind the brand?

Hyfe has three founders with complementary backgrounds and skill sets. I (Joe) am a data scientist with a background in public health, epidemiology, and economics. Originally from Florida, I’ve spent most of the last decade living abroad working on development and research projects at the crossroads of cutting-edge technology and public health. Paul is a software developer, product executive and entrepreneur focused on pragmatic approaches to integrate new technologies into existing businesses and the creation of scalable business models. He is handling software and engineering. Lulian is a serial entrepreneur with a background in exponential technology and impact at scale. He has more than 20 years experience in building and managing high-performing teams in challenging to extreme operational environments. At Hyfe he is focusing on venture building, fundraising and strategy. 

What the three founders share is the insight that, while technology has redefined every aspect of modern life - from shopping to social interaction to dating - healthcare has remained painfully stagnant in its adoption. 

There’s always a lightbulb moment before the beginning of a new venture. What was that moment for you?

The fundamental idea behind Hyfe - the idea that diseases have an acoustic signature that can be detected, quantified, and used for disease diagnosis and patient monitoring - is nothing new. In fact, doctors have been listening to the sounds of their patients for centuries. It’s just that working in the AI and the medicine space gave us a privileged perspective, where we could see that what doctors were asking for (the ability to track their patients remotely, the ability to detect epidemics quickly, the ability to scale up expensive, cumbersome, slow diagnostic services) was something that AI, if properly engineered, could do. 

In terms of specific moments, I think there are two: the first was early 2020, just as the world was confronted by a global respiratory pandemic in which coughing was an important symptom. This is when, with encouragement from a network of advisers, Hyfe has started. The second was slightly before that, when I was living and working in Nepal, and walking through the TB diagnostic process with a patient. The guy was sick - feverish, losing weight, and coughing up blood - and he had to leave his family and livestock, walk for days to a clinic, and then wait around for a truant healthcare worker to show up, only to be assessed based on self-reported symptoms and then sent him to take medication, with no way of monitoring whether his symptoms were responding positively to that medication or not. I saw that and I thought “we’ve got to be able to do better than this”.   

Tell us about your experience prior to launching your business?

The diversity of our founders’ background is one of Hyfe’s strength. Both Iulian and I have a background in public health, but from slightly different angles; my experience is more on the epidemiological and data side of things, where as Iulian comes more from the world of operations, interventions, and delivery. In fact we met in India years ago, working on TB. Paul brings to the team a technical and product vision that neither of us possesses, having built startups.

What is your day-to-day role with the company?

I work at the crossroads of (a) coordinating the dozen or so research projects around the world that are using Hyfe to track and analyze cough among their study participants and (b) leading our AI team so that the models we use are robust, low-bias, and clinically actionable. This means about 50% of my time is writing code and the other 50% is coordinating, communicating, brainstorming, planning, etc. Code-writing is mostly in Python and R; the non-code stuff is generally via video calls and interactive docs since we’re a fully remote company.

What inspired you to launch your business and what is the end goal?

I was inspired by the mismatch between (a) the amazing technology that has emerged in the last few years and (b) the almost total lack of technology in many healthcare settings around the world. That is, it seemed to me that the most powerful algorithms and methods were being used to solve problems that didn’t actually exist (like how to keep people’s eyeballs glued to a social media feed a few milliseconds longer) whereas the problems that really mattered (like how to prevent pandemics, diagnose disease, and make healthcare more inclusive) were largely being neglected by new technologies, particularly in the developing world. So, we launched this business to bridge the gap: to bring the most cutting-edge tech to the most important problems, and thereby make the world a better, more equitable, healthier, place.

How do you set yourself apart from other businesses in your industry?

Healthcare - and public health in particular - have an unfortunate reputation of being rather slow-moving. The complexity of the topics and the academic bias of many of the people involved in the industry often lead to overengineering. At Hyfe, from the beginning, we have strived to be user-centric and fast-moving. We launched a very basic version of our app within days after founding and since then have pivoted several times and changed consistently based on user feedback. This user-centric ethos is one of our biggest strengths.

Technically, we are also different in how we approach the ML problem we are solving. We are focusing on “real world”, naturally occurring sounds. This approach is unique in our field, where most teams build around datasets generated in the laboratory, usually dominated by “provoked/ elicited/voluntary” sounds. Our approach is much harder to pull off, but scales much better and also ensures high-quality data for our models to train on, along with user insights that will eventually convert into a better experience.

What’s the biggest lesson you’ve learned so far as an entrepreneur?

There are too many lessons to list! But the main ones are:

• Learn to trust others. I’ve found that micromanaging people is draining and time-consuming for the person doing the micromanaging, and demeaning and restrictive for the person being micromanaged. Everyone is better off when we trust, and allow others to flourish without too much prodding.

• Test assumptions quickly. We want to market with a very premature product, and this was one of the best decisions we ever made. Getting your assumptions off the whiteboard and out into the wild means that you de-bias your product quickly, and don’t build something that nobody wants.

• Be user-centric. There is a constant temptation to build something that investors, or scientists, or engineers want, and this is a mistake. Ultimately, Hyfe is a tool for people, and we need to make sure that people want it. 

• Adapt quickly to feedback. This one goes without saying, but it’s worth saying for emphasis.

• Don’t over-design/ keep iterating. We’ve benefited immensely from an attitude of “let’s just try it out and see”. It means that we don’t waste too much time imagining scenarios that never come to fruition, and it also means that we accumulate a ton of experience and feedback quickly.

With all the success stories around entrepreneurship and how innovative people have to be to take the leap. How do you think you’ve innovated your sector and why?

Healthcare has been lagging behind other sectors in allowing technology to transform/ improve the experience. This is particularly obvious at the interface, and it’s particularly obvious in the developing world. We are part of a larger “wave”/ trend of companies trying to change that. But I think we’re unique in that we know enough about healthcare to understand that the most successful innovations don’t replace the system, but rather work within the system. That is why we’re exploring partnerships with important players in the global health space; large research institutions, governments, etc.

What are your thoughts on failure?

Both scientific progress and entrepreneurship are built on the ability to learn from failure. We are constantly testing assumptions in the real world (i.e. optimizing for fast failure) and building on top of learnings. There is no innovation without a willingness to fail. That said, it’s worth differentiating between micro failures and macro failures: micro failures (an unsatisfied client, a bug in the platform, an incorrect analysis) are inevitable and extremely useful - and, by virtue of the lessons they teach - they prevent macro-failures. So, my thoughts are fail small and fail often.

What was the journey like when you decided to raise funding for your startup? And what tips would you give to early-stage founders getting ready to take the same path?

We were lucky at Hyfe that we were not forced to fundraise very early - we succeeded to unlock a bit of revenue early on, and we managed to build quite a bit of the product between the founders. This means that when we did fundraise we were able to do so without having to part with too much equity. This is an important lesson we have learned in previous ventures - don’t fundraise too early and don’t give away too much equity early on. To others on this path, I would advise them to build as much as they can themselves, first, before taking on outside people and outside capital.

What plans do you have for Hyfe over the next two years?

The next big challenge for us as we are growing is to increase our regulatory profile. This means that we will have to partially transition from a fast-moving deep-tech, consumer-driven team to a more typical healthcare profile, with better documentation and a realistic regulatory roadmap. 

How do you believe the evolution of tech will affect your industry over the next 10 years?

Healthcare will change radically over the next 10 years. The macro trends are there: mobile-first, data-driven healthcare, focus on prevention, radical shifts in business models. On top of it, technologies that are frontier now - AI, blockchain - will all mature and become mainstream.   Ten years ago, almost nobody had a smartphone. Now, almost everyone does. And that trend, the trend of people carrying around internet-connected microphones in their pocket, is not going to reverse any time soon; on the contrary, it’s going to solidify. There is a huge opportunity in acoustic epidemiology and those who know how to capture and process the mountains of data that this opportunity presents stand to benefit enormously.

What advice would you give to early-stage founders wanting to harness the power of technology to create a positive impact?

The most important thing is to actually take the leap. Historically, commercial models with a strong impact profile have been hard to fund and hard to scale. But there is a sea-change happening at the moment. We have an entire generation coming of age who truly care about impact and purpose, and this generation is slowly changing the way consumers act, but they also become investors, shareholders, LPs, and GPs in Venture funds, policymakers. So the time is ripe to build bold, impact-first models. It will take a lot of experimentation and I believe that the models that will truly scale will be built from first principles.