Interview with DeHealth’s Advisor: Viacheslav Kovalevskiy

16 Aug, 2022
14:43 min read

Viacheslav, hello. Please, tell a little bit about yourself to make DeHealth’s community more familiar with your professional background.
Hi. My name is Viacheslav Kovalevskiy, in short it’s Slava. And before we start I do need to offer a disclaimer — I’m here as myself and do not represent my employer — current or previous. Below are only my personal views and opinion. So, as I like to joke — I’m a podcaster and in my free time I’m an engineering manager. Or vice versa — you never know. Currently I’m working as Senior Manager in META, formerly Facebook, where I have built an org that is on different aspects of developers experience of anyone contributing to PyTorch. In fact, if you ever contributed anything to PyTorch, the chances are somebody from my org responded to you. We optimize the throughput of your changes, my org goal is to make sure that you have as few obstacles as possible when landing your PR. Before that I worked in Google, specifically Google Cloud, where I created an org from scratch which was in charge, for a huge part, of what used to be known as GCP Cloud AI Platform, now — GCP Vertex AI. Specifically, my org was the one Behind things like:

  • GCP Vertex AI Workbench
  • Deep Learning Environments
  • TensorFlow Enterprise
  • Managed TensorBoard

Apart from all that I also invest in startups, probably already around seven. And I’m in advisory boards of several of them, including DeHealth. And last but not least, I am founder of a startup that is developing new teaching methods that allows to scale high quality education and made it universally accessible, Kovalevskyi.Academy.

How did you decide to come closer to DeHealth — from technical advisor to DHLT token holder and finally — technical partner in DeHealth AI?
It’s an interesting question, because I knew Denys [Tsvaig, DeHealth’s Co-Founder, CEO & CTO] for a long time, and the question is in fact — how did I meet him? I really don’t remember. We were introduced to each other, stayed in touch, I paid attention to what he was creating… and one day …

So why did DeHealth interest me? Well, if you look purely from an AI perspective, currently in the world we are moving to a point where it’s much easier to get enough computing capacity to train as complex models as you need. There are a lot of types of models to solve almost any task you have. And what actually everyone starts realizing is that the key is data. It doesn’t matter if you have a powerful model — if you don’t have data to train that model — that’s it, without the data it is the end for you. The world is slowly moving from ‘How we can find the model that can draw a picture or predict cancer’ to ‘How we can find tons of annotated images to train models on these images’? The size and the quality of the data — that’s what matters.

DeHealth has quite a unique position because they have access to tons of healthcare data from countries which allow legally collecting that data. Here’s where an important part comes to play. DeHealth collects and poses data about people. People are all the same everywhere. It doesn’t matter which country you found that data — your model theoretically can be used everywhere. Even though there are still challenges in how to capitalize on that data, how to create these models, how to create this pass from the data to success, in my opinion DeHealth is technically half way there — DeHealth does have data.

That’s why I personally made a decision to invest in DeHealth. But, keep in mind, my opinion should never be treated as investment advice, investing via tokens is always the riskiest possible type of investment.

How does your previous experience help you now in creating DeHealth AI? What will it be in the end?
My main experience is — and always was — in building AI platforms and AI pipelines, i. e. taking an AI model built by different geniuses in this world and making from them a production service that generates money. And all these are covered by fancy names lsuch as AI platform or MLOps. To some extent this is more or less exactly what DeHealth needs today.

If we take a look at where DeHealth started — it was a telemedicine company that had access to data and tried to figure out how it could be useful to the world. It began with suggesting business strategy directions from DeHealth’s advisors from pharmaceutical enterprises as well as from merging with MIS in Eastern Europe which settled the basis for thousands of health records. With this data were many-many options of how we can proceed, depending on different ideas which were experienced. We can, for example, create a model which can ‘communicate’ with the user, learn more about the person, and finally, suggest (because in many countries you can’t diagnose — you are allowed only to suggest) about potential problems with their health and advise them to visit a certain doctor. This is one way — directly from the business to the consumer.

Another way would be from business to business, e.g. when DeHealth provides services to clinics. One of a clinic’s biggest problems is that they do a preliminary review of a patient, figure out the initial stages of the problem they have, and suggest the correct medical professional for them to visit. This is a huge problem, for example, in the USA: you need to spend several days before an appointment with a primary physician and then have to wait additional days to finally receive treatment. Medical insurances spend millions of dollars for that useless part of the chain. It can be easily removed if we have semi-automated way for clinics to access their patients’ data and directly suggest which particular specialist should be visited.

Finally one more option — and for me it looks most undervalued — is reselling the access to the data. The data is completely anonymized and DeHealth has this unique opportunity to provide other researchers with access to data for money. This is an interesting movement in the AI world right now when you allow someone else’s model to train on your data without giving the data away. It’s called ‘federation learning’, it’s quite challenging to implement, but if it’s implemented right, it might work.

Those are the directions which I consider the project can go and the team is exploring all of them.

Presently, or unfortunately, the healthcare market is way behind that level of sophistication and efficiency to what it should already be, and the reason is mostly because of legislation issues. In the USA they say no one should even attempt to enter this market because rumor has it even companies like Apple and Microsoft have tried to build their custom onside medical providers and failed. This is a good example that even big money are not able to push this market forward. This is mostly why AI has still not entered the market.

Another problem is that in many cases, even if AI has entered the market you can’t follow it even if the model is successful — the nature of current AI model is that you can’t explain why the model made a particular decision or why the model stated a particular diagnosis. It means you can’t use it because, if later the diagnosis turns out to be wrong, the patient could go to court and you’ll have to be able to explain why the model made that particular decision.

This is one of the many examples particularly preventing AI to enter the healthcare market. However, there are several ways to work around these issues, and I mean legal actions. The first is to start from developing countries or countries where healthcare is not so legalized, where there are not so many restrictions. A lot of western based companies usually just don’t know the market, how to enter Eastern Europe for example. Therefore, they are afraid of it and consequently loosing an opportunity to start something new. Let’s take Ukraine as an example and imagine we can integrate AI to do initial diagnostic in many facilities around the country. Europe and the USA will see that success story, will be forced to get their act together and start implementing a system that works so well. That approach might work because, as we see, it’s impossible to start that revolution directly from Western markets.

So, when somebody finds the way to penetrate that market from the back door, from Eastern countries, he will most likely succeed.

As we’re talking, you’re in Silicon Valley. Please, share your views on what’s going on there, the overall mood, aspirations, anticipations?
It’s quite interesting what’s going on in Silicon Valley, especially these days. We’re entering a crisis, a crisis of the startup industry in which everyone is slowly pausing hiring because of macroeconomic conditions. But in terms of startups, if you look for the last 5 years, the amount of money from VC was huge, enormous. I’ve seen people who are raising millions of dollars just based on an idea and slide deck — nothing else. At some point everyone realizes this model of investment — instead of investing in sharks you can invest in 100 startups and — as it used to be — you’ll find one unicorn. Because of constant growth of the income of people who live in Silicon Valley, a lot of them were paid with the shares of the companies they work in and now we have a lot of people with tons of capital, especially among those who came here in 2002–2005: they accumulated their ‘tesla’ or ‘google’ shares, never selling them to this point, and become venture capitalists.

What does this mean? There was much more money around than good ideas, so now we have tons of overvalued startups, far more than can actually produce. This leads to two problems. Firstly, when companies need more money they won’t be able to raise it because they are already valued too high and no one will value them higher when they need to raise more money. We can already see a number of startups who 9 months ago declared how much cash they had and now have started laying off people.

The second problem. As we’re entering this crisis, inverstors will probably start doing due diligence and evaluate companies better. Therefore, in a year or so we’ll see less startups popping up.

Interestingly, exactly what’s happened in Silicon Valley is starting to happen in India. Local engineers are at the point that Silicon Valley used to be in 2002–2003 when they had already earned enough money to start thinking about where they can invest it.

Apart from working in Facebook and been advisory in DeHealth you run Kovalevskyi Academy. What was the reason to set-up this online platform as well as the eponymous Youtube channel?
This is a very small startup. My whole life, before I started my professional career, I was teaching students. I loved that. One day I met the founder of hexlet.io, a Russian-language Coursera analogue and we started to collaborate. Java courses were the first ones I taught. At some point we parted ways and I realized there is a problem in the educational market. And the problem is quite interesting. The classic learning process presumes you attend lectures where teachers tell you theory, and then you have practice. But lectures are the most useless part of the process, and I’m stating this because in the world of Youtube you can ‘google’ any topic and there is tons of information. There is zero value for teachers to spend time on theory. The most important thing is practice because it shows how to solve real world problems. So what we did was create a new way of teaching people. We actually removed theory and teachers completely. We give students real problems that can be solved, by them, on their own. For example, let’s imagine you want to become a programmer. Your first task will be to create a certain program. With the group’s help you can google how to do this. And, even if somebody within the group is struggling with that task, the group will help him. The next day will be another task and so on. With this set of tasks we help the student to pass the road from complete ignorance in coding to their first interview.

This approach give us an opportunity to scale high level education, because scaling is always about approaching highly experienced teachers. This is the biggest challenge. But when you remove the teachers from the picture you can suddenly scale it. You just need a person to facilitate the group — that’s all.

Please, share your thoughts on cryptocurrencies and the tokenization trend.
First of all, let’s separate cryptocurrencies and tokens. Tokens are the tool to capitalize your work, to raise money. As a token issuer you can figure out what the value will be behind the token. In DeHealth, for example, they have access to the data, build the platform that will allow you to use the token [DHLT], and buy access to particular services and/or data. Now you have a value assigned to the token. You can actually start to exchange a token for money and start trading it. I don’t want to give anyone investment advise because tokens are still the most risky way of investing. However, I do believe in tokens and tokenization as a way of raising money and DeHealth has data behind it [DHLT] so I think they are raising money in a right way.

However, I don’t believe in cryptocurrencies. There is no value. As an investor I do invest in crypto but I have a policy of 80–20: 80% I’d invest in the most safe things such as government bonds from the USA because I do believe the USA wont go under while I’m still alive, and the other 20% — in the craziest things, like crypto for example, with full understanding that it can go to zero. It’s always a tail investment for me.

What do you personally think about Metaverse as the next step of people engagement? Perhaps you can share your thoughts on the development of the metaverse in terms of the healthcare industry?
I’m so far away from that team in META which develop VR and Metaverse and don’t have any insights, but even if I had I wouldn’t share them. So, I can only share my personal thoughts on that topic. Let’s first define the Metaverse. I do believe in it but I do need to put an asterisk. I’d define the Metaverse as a VR and AR world in which you have some way of creating and adjusting that reality and some sort of running economy when you can sell, buy things, etc. The reason why I believe in it, and can tell you that this is the future, is quite simple. If you’ve never tried to use Oculus from META — I highly advise trying it, because the moment you try VR specifically with Oculus, which has already reached minimum quality level needed in order to realise how it can work in the future, you immediately realise this will be the next way of communication. We used to communicate by phone, then Skype, now we have ZOOM etc. And VR will be way more convenient. The only thing is that Oculus is a heavy device right now. Once it becomes a glass that you put on your eyes it’ll be in everyone’s room.

So, yes — the Metaverse is the future, but when we will be there — who knows. Metaverse is a buzz word itself. What I described as communication opportunities — this is a specific problem which has actually been solved. Healthcare will be improved with Metaverse for sure — as soon as Skype — things will be replaced with Metaverse and this will be the place to meet your doctor. But there is still a bigger niche and we’re like 20–30 years affront tectonic shifts.

In your opinion, what sets DeHealth apart from other similar ventures?
There are several things. First — a lot of startups go like ‘we’re going to cure diabetics’ or ’we’re going to cure cancer’ while they actually have no data for that claim. DeHealth has data. You need to try very hard to fail when you have data.

The second point which makes DeHealth interesting for investors like me, who are searching high risk-high reward initiatives, is that tokens which you can buy today, will be high in value tomorrow. This means that the price of a token will grow exponentially the moment the data’s selling and buying options are available. You have a unique opportunity to invest now and get these disproportional results later, when the product is actually delivered.

So, the investment opportunity is there and the data is there. Now the question is to execute possible strategies. . DeHealth has a quite experienced team, I do believe that they have a high chance to succeed.

The data. The team. The tokenization mechanic. These three quite unique features are the key.

If you have questions about DeHealth, you can ask them directly to the community and CEO on Telegram . 

Related Articles

The biggest medical data theft cases
Article

The biggest medical data theft cases

Assaults on data on secured networks jeopardize every individual’s privacy, but attacks on medical d…

24 Sep, 2022
09:40 min read
DeHealth DHLT Network Integrates Chainlink Keepers and Price Feeds To Help Power Medical Data-Based App
Partners
Technologies

DeHealth DHLT Network Integrates Chainlink Keepers and Price Feeds To Help Power Medical Data-Based App

We’re excited to announce that DeHealth, a dApp for managing and monetizing anonymized medical data,…

15 Sep, 2022
05:27 min read
DHLT is listed on BitMart

DHLT is listed on BitMart

On September 08, 2022 world’s leading digital asset exchange BitMart started DeHealth native token D…

8 Sep, 2022
39 seconds read

You subscribed successfully

You have successfully subscribed, we will be one of the first to inform you about our significant updates.


Subscribe to our social media

Let's subscribe

Be the first to know about our news and updates


Subscribe to our social media