Founders of Karfidov Lab - about robots, AI implementation and IT staff drain

Alexey Karfidov - Co-founder and General Designer of the Karfidov Lab company. Studied at the department

metallurgy of steel and ferroalloys NUST MISIS, inpostgraduate study at MSTU Bauman. Prior to Karfidov Lab, he taught engineering graphics and worked in a government structure, headed the design department. In 2014 he founded Karfidov Lab with Dmitry Vasiliev. He is also a co-founder of a medical device subsidiary.

Dmitry Vasiliev - co-founder, director of company developmentKarfidov Lab, Medical Engineering and Komarik autoinjector. Graduate of NUST MISIS, master's degree in the joint program of the Skolkovo Institute of Science and Technology and MIT (Massachusetts Institute of Technology). Since 2014, co-owner of the engineering company Karfidov Lab. In five years, he has made the company a leader in the industrial design and science-intensive design of innovative devices with a turnover of more than 80 million rubles. He co-founded several spin-off companies, attracted investors to a project to create medical equipment. Expert in Product Design and Development, end-to-end design and development management, and manufacturing technology.

"Karfidov Lab" - design bureau for engineering design anddevelopment of technological products and their prototypes. Founded in 2010 as an unofficial student movement at NUST MISIS. Since its inception, the organization has participated in the development of industrial and scientific projects of the institute, twice won prizes at the CAD-OLYMP Olympiad.

In September 2015, the company received the status of the Skolkovo collective use center.

At the moment the company is workingmainly in the fields of instrumentation, robotics, medicine and transport. In the course of the work, such things as: torch body for the 2014 Winter Olympics, a helmet for digitizing actors' facial expressions, an exoskeleton, an underwater drone, a thickness gauge, a robot courier, radiation therapy and brachiotherapy complexes and many other products and complexes were developed and noted by the press. ... In 2020, the annual turnover of Karfidov Lab amounted to 100 million rubles.

"AI will predict which pizzas will be ordered today"

- AI capabilities are expanding rapidly. Experts saythat by 2025 it will be used in 90% of applications. What is the reason for this growth?

Alexei: The idea behind the widespread adoption of AI is tothe mobile application was able to predict some of the human actions. On the one hand, the user can independently define future actions. On the other hand, many applications are now customer-focused. The server processes data from multiple users who access it. And based on this information, the application automates many actions using various models.

Main tendencies in app development in 2020:

  • applications that create virtual and augmented reality in photos, videos and even maps;
  • using neural networks to work with video and photos;
  • cloud services that store information regardless of a specific device and work with huge amounts of data;
  • systems of electronic payments, as well as protection of funds, information and documents are being improved;
  • chat bots, especially widely used in online stores, taxi services and delivery services;
  • everything related to distance learning, training and work;
  • smart watches, rings, bracelets and many other portable devices that help to monitor health and plan time.

Let's say a company sells pizza.With each new order, she performs some actions. Using deep analytics, you can create a coherent predictive statistical model that will predict, for example, which pizzas will be ordered today and which ingredients will be required for them. The share of human participation will be less and less. He will rather act as an operator, support and agree on ideas, take into account recommendations and comments from the system.

Many functions will be performed internallyapplications. Fewer items of technology will be used. Analytics and data processing is based on information that is transferred from external hardware to a device or mobile application. Now, not separate technical subsystems and complexes are responsible for analytics, but applications. This is required in order to process all data streams quickly and efficiently, to give feedback, you need a permanent connection to the Internet.

Avoiding the human factor reduces the cost of wages and other support. This means that there is a good synergy between business and the development of artificial intelligence.

How companies should protect data

- How to secure your data in mobile applications?

Alexei: Data protection decisions must be madeinitially, when designing the very architecture of the application and its "binding", so that later the data is not located in an open environment. It is necessary to consider in advance the options and mechanisms for data encryption. Now this is a must. At the same time, the concept and architecture of the application must comply with the legislative and legal basis of the territory in which it is created.

Dmitriy: Several years ago, Europe adopted a law onprotection of personal information. This was a major legal shift. Everyone had to revise the policy of storing and processing any information, including paper. Russian companies doing business with Europeans were also forced to adjust, since the new law also affects counterparties. Most small and medium-sized Russian companies made the required minimum - notification of how the process of storing and processing information was built and made a notification to users.

Top 5 leaks information in recent years:

  1. Canva, an imaging site with 139 million accounts.
  2. Evite, an online invitation company, has 101 million email addresses, as well as usernames and dates of birth.
  3. 500px, made for professional photographers, leaked personal data of 15 million people.
  4. Quest Diagnostics Laboratory - 11.9 million customer data.
  5. Lab Crop, a laboratory affected by the same attack as the previous one - access to personal data of 7.7 million users.

Most sites now refer directly tousers to agreements regarding cookies or data collection. And upon request, if I'm not mistaken, the company is obliged to provide a document on how it handles any data: what it collects, from which users, how the data is used, where it is stored and how it is protected.

Now, when concluding a deal with a European company, one of the first questions to be asked is: "Tell us how you will protect and preserve our confidential data?"

“We were looking for a problem in the hardware, but there was simply not enough money”

- What are the general features of the development of new technologies or smart devices? What difficulties can arise?

Alexei: First of all, at least partially, but a new element base is used. For example, a new chip or specific operating features.

I'll tell you one funny example about cyclinglocks. When we made the first few test boards and assembled several ready-made devices, we could not connect to the network for about seven days: the connection simply did not go through. We rechecked everything several times, and everything seemed to be done correctly both from the server side and from the device side. The most annoying thing is that for the first couple of days everything turned on and worked well, and then suddenly it stopped. When you work with something new, a malfunction causes severe panic and you are ready for a lot to make it work. The reason was found out after about a week and turned out to be very funny. There was very little money on the SIM-device, about 50 rubles, and they ran out. And we were looking for a solution to the issue in the "hardware", architecture, re-soldering something.

Secondly, the use of a new element basehighly dependent on technical support from the person who produces it. For example, when integrating with mobile operators, many issues have to be coordinated. Let's say, from our side, everything is done well, but the device does not connect to the MTS. We have to interact with MTS support services for a long time.

There are also purely technical aspects:it is not known what the heat will be, what kind of piping, general technical issues common to all devices. The exact data in terms of power consumption is often only known when the first sufficiently long tests are carried out. Only after that do you understand which resource can be realized with the help of a certain element base.

Dmitriy: First, we must understand who we are making this device for, what it should do, what tasks to solve, what data and why to collect. This will affect the technology itself.

Second, working with data for AI and neuralnetworking can take longer than classic software tasks because data needs to be collected, prepared, described, and explored. The quality of the data also varies. There may be some anomalies or missing values. They need to be selected, cleaned, if necessary, generated new data, integrated, formatted. This is a huge layer of work that lies in the field of technology preparation. On the other hand, software tasks are still software tasks, for their solution you need a PC and software, and they are often solved locally. At the same time, to solve problems in the field of instrumentation, design bureaus, pilot plants, new or specific technologies, materials, and various specialists are needed.

"We have to hire self-taught people"

- One of the important conditions for the development of new technologies is the educational base. What is in Russia today in this area?

Alexei: The question is very acute and always topical.Because the ministry of education in any country does not always easily adapt to some new movements. It turns out that specialists in a new area are urgently needed, but there is no education in this area, since the ministries have not yet had time to develop and implement its plan. Then you have to hire self-taught people who are interested in understanding themselves.

Most Popular sites for online learning:

  • Coursera is a rare programmer who hasn't tried to learn a new language or library on this site with hundreds of different courses.
  • EdX is an international portal maintained by Harvard University and MIT.
  • Udacity - Developed by a Stanford professor, you can find both paid and free IT courses.
  • Codeacademy is an opportunity for interactive learning from scratch in programming languages ​​and web design.
  • Canvas Network allows you to find free courses not only from universities, but also from leading IT companies.

Dmitriy: The educational system definitely lacks flexibility.What students are taught was important 5-10 years ago. These training programs are extremely clumsy, complex, sometimes we see in lectures and seminars what is no longer needed at all. Many students come to us and get really necessary knowledge exclusively in practice.

Fortunately, there are exceptions.I studied at MIT and the approach is much more flexible there. Let's say the Computer Science direction consists of dozens of courses that are developed by teachers, depending on what is currently relevant. I can quite imagine that tomorrow there will be a narrow course on the development of autonomous robots in connection with the pandemic. People are not allowed to go out, and someone has to deliver everything to them. I would like such shifts to take place everywhere in our education.

We are given very good fundamental knowledge, we are in no way inferior in theory to the Americans and Europeans. But as far as applied things are concerned, there are serious failures.

- So it turns out that no theory can replace real practice?

Dmitriy: Here's another problem.Almost all of our subjects are taught more theoretically. Education abroad is structured in such a way that even if there is some theory, then later, during the semester, it is necessary to do some kind of team project. Students do something with their hands, write a program or software, make a prototype, where they will test in practice what they were told in theory.

In addition, now there are many opportunities for obtainingpractical education. You can find a lot of open and even free courses on the Internet. Therefore, the focus is shifting more towards online education. You open a page with courses in computer science and you see more than a hundred different subjects: a course on parallel computing, cognitive robotics, machine vision, and others. We do not have such narrowly focused subjects in higher education, and indeed in the compulsory educational system. The usual programming is in progress and that's it.

Alexei: I draw, teach young guys to model inCAD (computer-aided design system - "Hi-tech"). And I noticed that, for example, in comparison with 2008, when I was studying, now students think much faster, they grasp everything on the fly. 12 years ago, students were given tasks at the Olympiads, they solved with grief in half, puffed and said: "Lord, how difficult it is!" - and now, for the sake of experiment, I gave it to one group and said: "Guys, this is just an independent work, it must be done by next week." And they did. I was very surprised how easy they did it. And I came to the conclusion that technology is very important in the educational process.

Students who come now with a phonealready live, and the speed of mastering knowledge is also greatly increasing. For example, 10 years ago we said: “Guys, look at the GOST at home, in the evening” (GOST according to the rules for setting details in the drawings - “Hi-tech”). And indeed, someone was watching. Now, when you tell this to students, they all pick up the phone: “Okay, Google” - and in a second they open this GOST and immediately apply it. 10 years ago, we could teach material for one semester, and now everything is mastered in half that time.

"Data is the new oil"

- Artificial intelligence, and in general IT technologies, affect almost all areas of human life. In which of them, in your opinion, they are most in demand, and in which - on the contrary?

Dmitriy: It seems to me that the most interesting areas aretransport and mobility, industry, manufacturing and robotics; and computing and medicine. For example, according to reports, there are only five robots for every 10,000 people employed in Russia, while in Korea or Japan there are up to 600-700. That is, in this regard, we have a huge potential for introducing robots into the technological process.

At the same time, software will be in great demand,Machine Learning technologies. And necessarily the area of ​​the Internet of Things. But there are actually two areas of development. The first concerns our ordinary life, those objects that surround us. For example, an electric kettle, a lamppost on the street - thanks to being connected to the network and equipped with various sensors and sensors, they begin to generate data and adapt to the environment. And this collected information can be useful for business. For example, cafes will be able to understand where there are flows of people and when to expect visitors. That is, the business will be based on data. It is not for nothing that they say that data is the new oil.

The second is the Industrial Internet of Things:large enterprises, factories, industries that previously used the products of traditional, classical mechanical engineering, are equipped with a huge number of sensors, sensors. Machine tools and equipment are also becoming smart, digital. A large amount of data is collected, and based on this data, work is optimized, efficiency is increased.

Alexei: There is one general trend - all those areas wherehuman labor can be automated. This is an excellent platform for the development of AI, the development of new technical solutions that can be funded by medium and large companies, because they will be commercially interesting. This applies to autonomous robots, couriers and more. And I agree that the Internet of Things is a new, promising direction, which is now in trend. Devices are not connected anywhere with wires, they collect current local data from sensors, measuring transducers, and then transmit this information to the external environment. Typical examples: collect data from your meters to calculate your rent; garbage collection optimization: if the container is full, the application will send the filling status to the service that is engaged in cleaning.

- Are robots good or bad?

Dmitriy: This is inevitable, good and bad at the same time.You cannot look at robots from only one side. Of course, on the one hand, we deprive people of work, on the other hand, when a person loses his job, he can retrain and perform other tasks. You can call it another industrial revolution.

Alexei: In all countries of the world people sooner or laterexpress dissatisfaction with those who deprive them of their jobs. It should be remembered here that in the future, such changes will lead to the creation of high-tech jobs for the best results in one area or another. As a result, this will have a positive effect.

"You can bypass Russia as a logistics link"

- Over the past two years, Russia sank from 38th to 45th place in the IT development rating. What has changed in our country and why did such a fall occur?

Dmitriy: We disagree with this thesis, if there istype of IT security in various sectors of the economy. On the contrary, we feel that we are ahead of the rest, for example, in the financial sector, agriculture, manufacturing and medicine. Although there are no large Russian technology companies related to IT. In our country, the transition from technology to business is very difficult, there is a shortage of players in the market, it is difficult to compete with other countries in this regard. But this is not due to the underdevelopment of information technologies in Russia. Rather, we have problems with a market economy and the availability of finance.

Alexei: We ourselves as a company in the last two yearsmade a fairly strong leap and adjusted to modern realities in order to meet the requirements of the market. One can, of course, assume that everything is progressing even faster abroad, but we do not have such a feeling.

- What influence do the USA and China have on the development of IT technologies?

Dmitriy: Of course, they have a huge impact andare leaders, you just have to look at the market share of Asian and American companies. The capabilities of these companies extend not only within states, but also beyond their borders. We can say that they "hooked" the whole world to their technologies. Everything we use is either made in China or developed in the States. Of course, it is unlikely that tomorrow we will become world leaders, but we will definitely occupy our niche, albeit without the right to leadership. It is worth looking at foreign colleagues, adopting the best from them and going further on your own path.

To produce something, you need this somethingcome up with. It is not at all necessary to open production to make a profit, the main thing is that we have a development center. To set up production, you can bypass Russia as a logistic link. We know several Russian startups manufacturing goods in China and supplying them to Europe. It is necessary to allocate large funds for education and science. In our country, the percentage of research and development in the structure of GDP is about 1–2%, while in Northern Europe it is several times higher, although they are not world factories. It's also important to maintain creativity. Roughly speaking, it is necessary not only to build factories and factories, but also to provide financial support to people who come up with new and unique things. China, for example, now has a government program to support the creative industry.

Top 10 countries with advanced AI:

  1. USA - AI is used in military affairs, for processing data, creating more flexible systems.
  2. China - AI is mainly used for trading, in particular on Tencent and Alibaba servers.
  3. Great Britain - The British are using the capabilities of AI in medicine, biology, research and development, telecommunications.
  4. Canada - creates autonomous vehicles, equipment for the forestry and mining industries, nuclear power and hydraulic engineering, for geophysical research.
  5. India - actively develops and exports software.
  6. Israel is developing genetics, optics, agriculture, electronics, and the use of solar energy.
  7. France is AI in all possible forms.
  8. Germany - developments in robotics, speech and language technologies, intelligent visualization and modeling.
  9. Sweden - technologies to make life and work of people easier, for example, "talking heads" at airports, machines for sorting waste batteries using AI.
  10. Spain - technologies in agriculture.

Alexei: Our development in technology is hampered by problems,emerged after the collapse of the Soviet Union. The majority of electronic components are manufactured in China for a number of reasons. Fast delivery within the country, as well as in-house production, helps to speed up assembly and improve results in the future. At the same time, our country is experiencing difficulties in obtaining the same materials. So far, we are producing an insufficient amount of our own element base, which prevents Russia from developing faster.

- Is staff drain a problem?

Dmitriy: I would not say what it is.People need to be mobile. If someone wants to go somewhere - let him go, get the necessary knowledge, be realized. I personally had the opportunity to study in the United States, and upon returning to Russia, we created a technology company. I think this is the approach that needs to be promoted. The main thing is to create the best working conditions in our country. In the States, for example, a huge number of Chinese are studying, but they have no desire to move there permanently. They return to their homeland as soon as they finish their studies, because in China they have a huge number of opportunities for future development.

- How can Russia become a leader among technology companies and be able to compete with the United States and China?

Dmitriy: Quite a difficult question.Probably, you need to act quickly in order to catch up and, if possible, even get ahead. It is necessary to allocate a large amount of resources for education, new developments. Due to the lack of sufficient money, current Russian startups turn to Western investments and end up in the States or Europe, gradually flowing there from Russia, simply because everything is easier there with financing.

Alexei: It would be great if more smallenterprises, and especially - simplified bureaucratic procedures, the organization of the export of products. Due to my patriotic convictions, I believe that it is important to maintain technical competence in production. Now in Russia, by 2025–2028, it is planned to develop in the production of automobiles, aviation, and mobile technology. This is great and will increase production capacity in the future if IT keeps up. The creation of robotic autonomous complexes that can conduct their activities without human intervention is an additional cost to the economy of the entire country. The ability to create such complexes and bring them both to the domestic and foreign markets can have a positive effect on the number of high-tech jobs.

See also:

The first accurate map of the world was created. What's wrong with everyone else?

Infrared radiation from human hands was used for encryption

Uranus has received the status of the strangest planet in the solar system. Why?