AIME – Artificial Intelligence in Medical Epidemiology. Start-up of the week.

Last summer we heard a lot about zika as many athletes were withdrawing from Rio de Janeiro Olympic Games fearing this virus. Fortunately, there were no reported cases of anyone being infected. AIME and its platform capable of forecasting the exact geo location and date of outbreaks of infectious diseases like dengue or zika three months in advance probably had something to do with that.

Logo AIME - Artificial Intelligence in Medical Epidemiology

I met Dhesi Raja, co-founder and Chief Scientist of AIME Inc., in Singapore, as he was one of the ten Innovators Under 35 chosen to present a three-minute elevator pitch at EmTech Asia (you can read more about the highlights and best quotes of the event here).

In those three minutes, Dhesi managed to present their Silicon Valley based startup, their platform to predict dengue and zika outbreaks three months in advance with a 86.37% accuracy, their mobile app to chase breeding sites like they were Pokémon’s and their future plans to expand to other epidemic diseases, like diabetes or hypertension.

If he was able to do all that in three minutes, imagine what he did in a 30 minute interview!

We are not sure if our actions will save lives, but we are certainly sure inaction kills – Dhesi Raja (Chief Scientist of AIME)

 

Founders Startup        Process Startup        ecosystem country

Funding Startup        Future Startup        Advice Startup

 

A Doctor and a Computer scientist, the brains behind AIME

Dhesi Raja (Malaysia) and Rainier Mallol (Dominican Republic), a 32 year old Medical Doctor with a Masters in Public Health and a 25 year old Computer Engineer with a Masters in Networking and System Administration respectively, saw in each other the perfect combination to found AIME in 2015.

 

Co-founders AIME. Dhesi Raja and Rainier Mallol
Left: Rainier Mallol. Right: Dhesi Raja

 

Two guys with completely different backgrounds and from completely different parts of the world. Dhesi had already been doing research for 2.5 years on identifying the variables leading to dengue outbreaks when he met Rainier. He told me why they were such a great match:

“I realized that in public health we have failed to contain outbreaks. So if you see Ebola, we failed, talk about SARS, we failed, and recently zika, also we failed. The reason we failed is because there is no input of computer science in pandemic control programs.

You see a lot of science or innovation in finance, in logistics, in transportation, even in medical science like genomics, but if you talk about outbreaks of pandemic, you rarely see innovation there. We needed to bring public health to a different level, not at statistic prediction, but AI (Artificial Intelligence) prediction.

So I joined the competition for the Global Impact Competition in Malaysia , won the first prize there and got a scholarship from Google & ECM Libra to Singularity University in NASA. I presented my project there and that’s when I met Rainier Mallol (also a student at Singularity University) and he told me that we could make it big.”

 
Being from the Dominican Republic, Rainier was familiar with the dengue issue. When he knew about Dhesi’s research, he saw a big opportunity to input his knowledge in Artificial Intelligence and Machine Learning and make an accurate prediction model. The seed of AIME had already been planted.

 

 

From Silicon Valley, to the world

“We won the top 5 projects in Silicon Valley at the Singularity University, so we got our first grant, the FIX (Field Innovation Exchange Program) and went to Rio. When the Government of Rio saw our product they were happy, so they asked us to help them for the Rio Olympics, to prevent dengue and zika.

After helping the Government of Rio, we won the first prize in King’s College in London and also the Top 8 Young Health Innovators in Harvard. Because of that, the Government of Philippines heard about us and asked us to go there.

Now, we are already involved in six countries. Currently we are in three (Brazil, Philippines and Malaysia) and we are speaking with the other three (Dominican Republic, United States and India).

We’ve developed an autonomous system so every 23 seconds it tacks 270 variables autonomously, automatically. Rainier Mallol has created APIs (Application Programming Interface) for that.”

 
270 variables affecting the likelihood of dengue outbreaks! I couldn’t resist, I had to ask Dhesi which one had surprised him the most:

“Actually it’s thunderstorms. Because you see that with rainfall and wind direction, it makes sense. If there is more rain, there is more stagnant water, so there are more mosquitos. If wind is blowing to the West, there are more mosquitos in the west. But I was surprised to find that the higher the thunderstorm, the higher the number of breeding sites. I cannot answer why.”

 

AIME fighting dengue in Rio Janeiro

 

And what about the challenges they encountered to reach that expansion?

“The main challenge was to convince people to invest to protect the country. If you have a cough and cold and go see a clinical doctor, you are willing to pay the money because you know that you recover after that. But telling someone to pay for a disease that has not yet happened is not an easy task.”

 
Actually, a lot of money could be saved by investing in prevention. As is often said, prevention is better than cure!

In 2015 Brazil spent 1.1 to 1.2 billion US dollars to fight Dengue. That’s a lot of money! And zika has already cost the world more than 3 billion!

 

A mobile app to share its technology with everyone

So far, AIME had been a web-based app for government officials only. Whenever a government required its services, they would go there, input the country data and run the predictions.

A call from WEBE Community, a Telekom Malaysia subsidiary, changed things, and they started to work together on developing a mobile app, which will be launched later this month on the App Store (the launching ceremony will be held on April 18 in Kuala Lumpur).

“Dengue and zika are community diseases. So when the Government says that we cannot give the community the data and the app, because they will panic, we tell them that they must understand that if we do not share this data with the community there is no way the community will know that there is an outbreak.

It is really important for the community to know, to be able to clean the area, because it’s dengue, so it’s all about mosquitos. If you clean, you don’t get the disease. If you don’t clean, then you get the disease. Finally the Government agreed, so we created the app.”

 
It was one thing to create the app, but the second problem was how they could create stickiness, so people would want to use it like they use Facebook or Snapchat?

“We created a model like running for chasing Pokémon, but this one you chase for breeding sites. You find breeding sites, snap a photo and send it to us, then we will send it to the City Council to verify. The City Council will send people to check and if it is confirmed, you will receive free credits and data plan. So users are happy, they report the cases and the Government can go and clean the areas, stopping the disease.”

 

Pokémon Go

 

I was fortunate enough to see the AIME app to beat dengue before it is available in the App Store. It really looked like Google Maps, except for a piece of text below the map saying: “now you are currently not located in a dengue prone area”.

“Now it is offline, you could turn it on when you are travelling in Asia. If you enter into a dangerous area, the text will say that you are currently entering in a high risk or moderate risk area.”

 
Apart from the Pokémon Go like one, the app had other very cool functionalities:

You can ask AIME what are the possibilities of you having dengue. AIME will be tracking all your travel, so it will see all the high risk areas where you have been travelling. It will ask you a few questions about your symptoms and it will diagnose you whether you need to see a doctor or not.”

 

Singularity University

At this point, no doubt some of you may be wondering what this Singularity University where the project was born is. Located in the NASA Research Park in Silicon Valley (California, US), Singularity University annually offers an exclusive 10-week program for entrepreneurial talents. Being also a business incubator, one of the requirements to be accepted is to prove that your skills can have a big social impact.

“They only accept 3% of the applicants. What they do is try to solve global grand challenges. The Singularity says that in order to make a billion dollars you must solve a billion problems and impact a billion lives.

They train us in various challenges, so it is up to you whether you want to solve the global grand challenges of water, global grand challenges of health, or global grand challenges of environment. So that is when I decided to go for health and that was when I saw Rainier Mallol. There were only 60 students accepted during my alumnae.”

 

Dhesi Raja at Singularity University at NASA

 

In order to train people to have such a big social impact, they focus on changing people’s lineal thinking for exponential thinking. That’s the same kind of thinking behind the Moore’s Law that allowed the rapid expansion of computing by doubling the number of transistors in a chip every two years.

I was very interested in knowing what studying there was like. I imagined that being in such a selective institution, they had to feel a lot of pressure.

“There wasn’t pressure, it was very exciting because no one will look down on you if you have stupid crazy ideas. The problem with Asia is that when you present new ideas people think of the Return On Investment. But in Mountain View, in Silicon Valley, people don’t talk about money yet, people talk about how can your project impact lives or impact a billion people. Once you get that right the money will come in later.

We also had wonderful lectures from Google engineers, from Ray Kurzweil, one of the inventors and also the head of Google engineers, lectures from Guy Kawasaki, one of the few guys who started Apple with Steve Jobs. They were excellent mind blowing lectures.

And they are networking and connecting us. This is very important in a startup. Because you can have the best technology and the best mind-blowing idea in the world, but if no one connects you to the right client or the right person, it’s not going to be successful.

It is not only about investment, but to connecting with the right client. AIME has been able to self-sustain without a single investor. We have not been invested in by anyone. We have achieved our entire business model so far by subscription model by the client, so the client has been paying us, and we have been doing it successfully.”

 

Funding the Project

“Initial money was between me and Rainier. Because we had the brains, so we didn’t have to pay anyone. We built the prototype. Only when we tried to expand to Brazil, Singularity supported us with the FIX grant. Once we got there, the Government of Brazil sponsored it. Then came the Government of Philippines and currently the Government of Malaysia is paying us.”

 
I was interested in knowing if apart from the FIX grant they won in the contest, the Singularity University helped its students funding their projects:

They don’t fund us but they provide the tools. It is up to us to find those funds. They will be very simple, they will ask you whether you want to just grow your business based on your clients, you want a VC (Venture Capital) or you want an Angel Investor. Everything is at Singularity University. They can get the richest man there to listen to you if you want an investor.”

 
But Dhesi and Rainier weren’t interested in getting investment:

“We didn’t need that because I wanted to focus on the tech part, the Epidemiology, because not many people understand the Epidemiology part of disease prediction. There are many data scientists and computer scientists out there. Everyone is interested in creating a good mathematical model, but at the end of the day it is you testing down the field, because what you see in a simulation is very different from the real world.

Of course, I agree to having VCs and funding to increase your value, but I feel we are an academic startup, so it is a bit different. Uber can afford to have a wrong cab come to the wrong place. We can’t afford to do that. So when we say there is 84% that you might get zika, I bet you it is around 84%. We can’t afford to make mistakes in that.

 
I had read about a Google attempt of doing something similar of what Dhesi and his team are doing with AIME. It was called Google Flu Trends, and it provided estimates of influenza activity. Some years later they also launched a similar tool called Google Dengue Trends. Both services failed, apparently they prediction capability wasn’t good enough. A lack of epidemiology understanding?

“Correct. Now they have realized and are recruiting some epidemiologists to listen to them. But in this area you should get epidemiologists that are really interested in data science and tech. If not, it won’t work. Because doctors have they own mindset, they always think they are right. It is easy to work with computer geeks, but it is not an easy task to work with doctors. You need to be really humble, listen and argue.”

 

Google Flu Trends

 

Going back to the funding issue, Dhesi told me that they were exploring other possibilities before raising capital:

“We are speaking with one insurance company to predict possible heart disease, hypertension and diabetes. And we are also looking into another pharmaceutical company, they want us to predict for them where the best clinical sites are to do their study. We are also in the midst of talking to the Government of Florida to help them with zika.

So, after looking at this, we see that there is no need for us to raise capital yet. Of course, if there is a very good investor who is willing to invest… We had a few investors coming to us and talking to us, but the stake of the equity that they requested was too high. We don’t want someone to dictate us what to do.”

 

AIME’s vision for the future

When asking Dhesi about their vision for the future, I got the kind of answer I like to listen, a very ambitious one!

“We want to be the next Google of infections, diseases and health. What we want to do is that whenever there is something to do with diseases, people think of AIME.

For example, in case I want to plan a trip to Brazil, to Argentina, Chile, or whichever part of the world, I need to ask AIME.”

 

A disruptive piece of advice

Remember when you were a child and your mum and dad told you over and over to listen to older people as they knew more than you? Well, they might have been wrong:

Do not listen to experts, because experts are good at telling you why you cannot achieve your dreams. They are already an expert in that field, so they will tell you all the possibilities of why you cannot do this, you cannot do that.

Even Thomas Watson, founder and CEO of IBM said that in this world there are only five computers needed, but today every one of us have a computer. So I think we have to stop listening to experts and start following our dream and keep trying and trying. It is what Elon Musk said, as long as you follow a theory of reasoning, you’ll achieve your dream.”

 

Dhesi Raja at TEDx Kuala Lumpur

 

Thank you Dhesi for this eye-opener piece of advice! We’ll keep close track of AIME’s growth and expansion. Definitely, the world needs more start-ups like this one!

 

Dhesi Raja and Borja Bonet at EmTech Asia


From Singapore to Laos. I bet you don’t know much about what’s going on in the Lao start-up ecosystem. I was there and I had the chance to interview some of the local pioneers. Sign up below if you don’t want to miss what they told me! 😉

Borja Bonet

Hi! I'm Borja, and I'm so glad to be a part of this project and to be sharing ideas and experiences with all of you! Let's do it together! Welcome to Ambitious Tracks!

One thought on “AIME – Artificial Intelligence in Medical Epidemiology. Start-up of the week.

Leave a Reply

Your email address will not be published. Required fields are marked *