Fresh Perspective: About the Project
A large and diverse range of people study and work at HSE University. A Fresh Perspective series introduces some of the newest members of this extended community. Each brings a unique background, worldview and understanding of how a modern university should be organised. Regardless of whether you agree with them or not or find their ideas controversial, they offer a fresh perspective on HSE University.
How will the university spend the funds it has invested in its digital transformation? When will we see SmartLMS 1.0 and is it realistic to switch to a completely paperless educational process? Why does HSE need a supercomputer and when will artificial intelligence replace the university rector? HSE Senior Director for Digital Transformation Alexey Chukarin discusses these questions and more.
You came to HSE University in 2019. Why did you leave the City of Moscow IT Department and what sort of tempting offer did HSE make that convinced you to come?
I worked for eight years with the MITD, starting in 2011. This was a difficult and interesting time when we created many new products and solutions. The MITD team managed to transform the digital and electronic landscape for Muscovites. We created such tools as Active Citizen, Our City, a state services portal, and a municipal video surveillance system. And the creative phase of all these breakthrough projects came in late 2018.
After his re-election, (Moscow Mayor) Sergei Sobyanin set a new goal of making all of these solutions operate with the greatest possible degree of stability and quality within the next three years.
This was less interesting for me. This was when HSE asked if I would like to do almost the same thing for the university. They said I could create whatever I thought best, but to make HSE the digital leader across the entire educational landscape — and hopefully, beyond. That possibility was too good to pass up.
How has the IT department at HSE changed since your arrival? Were you depressed at first, and are you encouraged by the progress now?
When I was just starting to plan the sequence of steps to take, I looked at the services from the standpoint of the user because, in the first stage of immersion, I was more like a user myself. I looked at the HSE services for staff and students, at the ‘beautiful’ LMS and employees’ online personal accounts. I was a little upset, but I decided: ‘It’s not so bad: we’ll put everything in order quickly and by the New Year, everything will be just fine.’
Then I began ‘digging down’ through the services to the underlying system, and at this point I became depressed. The systems weren’t bad, but they were created more than 10 years ago. It is very difficult to integrate them with anything else or modernize them.
Here we came to the main task that took six months to solve — opening up the systems and getting them to work with each other. Wherever we succeeded, we set them aside for further development. But we will gradually replace those that are ‘set in stone’ while retaining everything we have learned in the process.
Then we saw the scale of the infrastructure problems — networks, servers and everything else, very old equipment that had not been repaired or upgraded for some time, and that had almost completely exhausted its service life. I was surprised that it had lasted so long.
At the same time, HSE had tried to maintain its own engineering services as long as possible, which was an unrealistic goal in the current competitive market. We are now in talks to try to adopt the IaaS (Infrastructure as a Service) model this year. This includes data centres, networks, and server facilities.
We could use all the funds allocated for the university’s digital transformation to purchase equipment, but that alone would not make us the digital leader as called for by the rector.
Now we come to the subject of money. HSE University has allocated 4 billion rubles for digital transformation over the next four years, 1 billion of which you will receive in 2020. This is an enormous sum for any Russian university, but is it enough to achieve your objectives?
Of course, we will do everything possible because the sums involved correspond to the designated task — although, after detailed calculations, my colleagues and I requested many times more money.
We have optimized everything possible. We have transferred every possible solution to the operational model and switched every possible outsourcing solution to insourcing, significantly reducing the final cost
According to my estimates, the funding shortage is not 100% now, but approximately 50%. We can find a way to manage with that: it is a typical management issue that we can resolve.
We all understand perfectly well — especially considering that we are at HSE — that it makes sense to organize investments in phases. We have a lot going on right now, much of which is happening simultaneously. But until we move to a new platform for education, to SmartLMS, it is useless to simultaneously launch a bunch of tracks with an open interface for developers who can create their own solutions, because then we would have to redo everything when we launch SmartLMS six months from now. This is the correct managerial approach and I am certain that if we happen to achieve our 2020 goals early, we will immediately begin working on the objectives for 2021.
We are increasing the digital staff because full-time employees are an indispensable resource that can solve a wide range of tasks at no additional expense.
Are you hiring more engineers and developers? You mentioned that it is more cost-effective to outsource some things and keep others in-house. Can you be more specific?
For example, everything related to know-how and competencies, the automation of educational activities or the management of internal HSE processes should remain in-house. This is valuable HSE practice and it makes no sense to outsource it.
Eighty per cent of all work that requires narrow specialization as well as one-time activities such as engineering support of data centres or software development according to pre-determined technical parameters will be outsourced because there is no reason to focus our entire technical team on such tasks.
That is why the digital group now forming in primarily focused on system and business analytics and project managers. We have a small group of programmers because there are internal systems that are easier and faster to fix or modify using our staff. We also have a large technical support pool.
Where do you get these people?
We managed to retain the Directorate of Information Technology (DIT) team, with very few exceptions. DIT is now part of the digital unit and focuses on the development and quality assurance of technical user support.
All other staff came from different places. We have almost finished forming a new team with people in all key areas. This includes individuals who have worked at HSE and specialists who came here with market experience. Although it was something of a quest, I managed to lure professionals from MITD, Innopraktiki of Moscow State University, Alfa Innovation, Sberbank, Systems, Tinkoff Bank and other leading digital companies to work on our interesting tasks.
We have assembled a team that is focused on results. No other approach can work. If a person comes to work at 10 in the morning and goes home at 7 in the evening, and also finds time to eat lunch, we definitely won’t achieve our objectives in under two years. As I tell all the newcomers: ‘With us, it’s frightening but fun.’
Will all this effort produce solutions that are unique to HSE University?
You have to look at it case by case. We are putting together HSE SmartLMS for the education process but, of course, we are using ready-made components to expedite the development process. We have also created many things and the market is full of solutions that provide for this or that function. For example, it is pointless to create a personal account from scratch because numerous ready-made solutions are available on the market: you can just choose whichever one you want and use it as a component. Most important is that the solution be compatible with the platform engine — 100% of which we developed ourselves.
The second major platform is back-office solutions for the support unit. We could have chosen any one of several different approaches, but my colleagues from the support unit and I decided we would use a single 1C platform on which we would deploy many different configurations: accounting, salaries, personnel, document management, migration registration, and so on. We will configure each element specifically for HSE, saving time on implementation.
Do you prioritize certain products, and if so, how?
We have about 15 products. These include supply unit products for finance and personnel. There is also a ‘properties’ product and a ‘universal corporate solutions’ product that includes all corporate systems such as electronic document management, CRM, and a unified calendar. However, four particular products are the highest priorities.
The first is the ‘education’ product — that is, the SmartLMS. This is the most important task before not only the digital unit but also the entire university. The number of students is growing rapidly, as is the number and complexity of programmes, and we need to manage all of it online. This makes the ‘education’ product the most important.
Second on the list, though not necessarily second by importance, is the ‘integration’ product. We need to ensure integration between internal and external systems. Otherwise, students will have personal accounts that look good but that contain fragmentary or incorrect personal information, instructors will see only pieces of the class schedule, and the training office manager will again wind up moving data between systems in semi-file mode. This makes integration important primarily for ensuring the quality of educational systems.
The third product is ‘information security.’ HSE University has faculties, labs, and departments devoted to information and computer security, but we are like a shoemaker without shoes . …Unlike in many other areas, no one at HSE has been working on information security. We came under active attack this summer, but the countermeasures we employed prevented the assailants from carrying out a single major plan.
Our team is working with experts to protect the university’s existing digital services and those we are creating from possible cyber threats.
Information security is one of the top priorities. Of course, we must still work out a million different things, but we have already managed to rebuild protection for the perimeter and to establish the necessary data protection.
Several information security projects carry a high priority: the introduction of a cloud-based digital signature that would simplify the signing of documents, regular audits of information system security, modernization of protection services for the network perimeter, anti-spam and phishing protection, anti-virus protection, and implementation of information security policies following relevant Russian laws.
The fourth product is ‘infrastructure.’ Without high-quality Wi-Fi, a fast network, and a well-functioning data centre, everything else has no meaning.
Do you have any tentative dates for launches, beta testing, and product introductions?
Of course. We will be rolling out closed testing of the MVP (minimum viable product) of the SmartLMS in March. It will be a working system, but with a limited set of courses for only some departments and without all the functions of the final product.
We plan to release SmartLMS 1.0 this September to start the next academic year with a new, large information system that will incorporate everything positive from our two existing systems and provide new digital educational services.
Of course, we will not completely replace our current servers by September 1, 2020, but we will integrate them into a single system. We will also remove the services that are most in demand, give them a new architecture, and reinsert them like building blocks.
How will you integrate all of those products into a single system?
From the standpoint of architecture, we decided to create a single system for working with data and a single corporate bus. This is a strategic system — a system of the Master Data Management (MDM) and Enterprise Service Bus (ESB) class.
I will try to explain briefly what this is. We have many different systems, each with its data model. The same person might appear as a student in one model, as an employee in another, and as a researcher in a third. Problems can arise when that person tries to bring them all together in some way. The MDM system allows us to gather all this data in a single structured storage area. We will not replace any of these systems’ data and, at the same time, we will preserve the keys marking how an individual went from being a student to a researcher. This will give us a single cohesive record of all such entities — people, campuses, legal entities, and documents all become centralized with links to different systems.
Rather than a data manager trying to organize all these processes, the ESB solution makes it possible for open APIs (program interfaces) to manage them at the level of services. Part of the reason is that when we open this API to everyone, it can implement any application, whether for a mobile device, desktop, or portal. This is not a breakthrough, but an advanced approach because the best practices in building corporate digital systems involve this type of centralized integration.
What about infrastructure?
We will switch over to the IaaS model in the first quarter to handle server storage of data and networks. We will radically change the quality of the final service, especially Wi-Fi, which will become seamless.
We have also completely restructured the work of our supercomputer complex and the digital unit now runs it. We invited an accomplished team from South Ural State University and created a supercomputer-modelling department. The team has managed to do more in its first three months of work than was accomplished in the previous 18 months.
Why does HSE need such a powerful computer?
Modern science cannot advance without supercomputers and many universities have them for this reason. The Skolkovo Institute of Science and Technology has a great machine and South Ural State University has three of them. Moscow State University, St. Petersburg Polytechnic University, Nizhny Novgorod State University, Tomsk State University, Ural Federal University, and others have them. This is not the exception, but the rule for modern universities. After all, the supercomputer is a tool for conducting scientific computations and performing complex research.
A number of HSE faculties are conducting very serious scientific research. They include the Faculty of Computer Science, whose skilled AI scientists use the supercomputer to create and verify new algorithms for training artificial neural networks. There are also the scientists of HSE’s Moscow Institute of Electronics and Mathematics who perform many interesting tasks in the natural sciences. And then there is the Faculty of Economic Sciences and the School of Sociology, that also requires enormous computing power to build various sociological models.
Our supercomputer is currently working at 72% of capacity, which is a relatively high indicator. Of course, we have enough reserve capacity to meet growing demand until the end of this year, but our scientists will soon begin vying for computing resources and asking us to enlarge the computer.
We achieved this 72% indicator fairly recently. Our colleagues who joined us from South Ural State University discovered a second, undocumented supercomputer management server with the help of which it was possible to increase the security and controllability of this complex tool substantially. They also found numerous errors, loops, and repeated use of the same resource in the task queuing system. It seemed as if our supercomputer was operating at 80%-90% of capacity when, in fact, it was only at 30%. We have now restructured all of that with the result that users can utilize many times more computing power on the same equipment without interfering with each other. The new employees ensured the stable operation of the supercomputer, which led to a rapid increase in the number of users and research projects carried out on the supercomputer.
Our supercomputer now enables us to solve a very wide range of problems and not only those connected with the university. We are now consulting with HSE Rector Yaroslav Kuzminov on how we can use our supercomputer for interesting research initiated in the ‘outside world.’
What is your planning horizon? Do you ever look beyond it?
We have three planning horizons. The first is operational and spans the current academic year. This is a university and the most important day is September 1. Our closest planning horizon is, therefore, September 1, 2020. The second horizon is September 1, 2021. My colleagues and I started working on changes in the technologies for supporting the admission campaign. Because many volunteers are involved, this process requires a simple set of tools that offer the greatest possible convenience. The third planning horizon is the three-year period.
If to look beyond these planning horizons, then mobile applications, artificial intelligence, and cloud solutions will become increasingly important. We will do more and more in the cloud each year.
And, of course, we also have educational solutions: simulators, trainers, gamification and mobile applications for education. We are now actively discussing closing the electronic cycle of student record books such that instructors will be able to use any electronic device to assign grades and sign the relevant forms with an electronic signature — and all of this will be posted to existing student records.
As a result, HSE will be the first Russian university to switch completely to paperless support of the educational process.
When will artificial intelligence replace the rector?