Welcome to CIO Central. I'm your host, Bill Miller, and I have today a very special guest with us, Theodore Prokarov, both CIO and chief technology officer for SpairCloud. SpareCloud is a subsidiary of Sparebank, one of Russia's largest financial institutions. Uh, welcome to the program this morning, Theodore.>> Hello, Bill. Hello, viewers. Hello, viewers. Hello, viewers. It's very nice to be here. >> Tell us a little bit about yourself, your background, and maybe specifically how you came to spare cloud and what part of this mission excited you. >> I'm coming from the uh IT in financial sector background. I was uhresponsible for a few roles in city group in Russia and finishing as the CIO for the corporate bank of city for a few years. Then I've switched to retail banking being a CIO of the settlement bank with a joint subsidiary of bear bank and BNB pariba. Then I've uh spent a year in Bear Bank uh being responsible for the IT team who is developing the and supporting online uh banking for the corporate clients and then I was uh pleased to join the Bearcloud team uh as a CTO being on the edge of the cloud computing machine learning and all the technologies uh that we have here. You clearly are an accomplished practitioner as a chief information officer, a CIO, but a lot of CIOS don't become CTO, chief technology officer, and indeed in this case, a true architect of a new computing environment. Um, can you explain that combination role for our listeners today? Well uh during my whole career I was approaching the uh IT uh and specifically banking IT from the different aspects. I was uh managing projects. I was managing operations. I was building data centers. I was uh optimizing insourcing and outsourcing the operational processes in different environments. And I think all of that uh helped me to develop in a uh person with a quite a broad view to look a few steps ahead to uh think about the technology advanced in general and uh also to look at it through a prism of the uh really uh pract practitioners point of view on how the particular technology can really help not just in marketing but in real life. So you get to be the business side of the house is the CIO understanding the nature and the economics and the services of what you're trying to build but you also get to be a little bit engineer technologist and architect. It sounds like that must be a fun way to spend your days. >> It's a very interesting environment to work in from both technology and processes and operational points of view. and being on the front edge of all that uh supercomput and uh cloud technology is extremely interesting but it's also very diverse starting from the uh basic Y services uh based on the virtual machines and virtual compute environments we also have some uh products uh like PS with uh building blocks uh which are managed services and uh our core competitive advantage is the MEL as a service which we do on our Christophari supercomputers >> and for ourlisteners here Christopher as I understand it is the first true super supercomput u built designed in Russia for commercial purposes and commercial users so how did you come todecide it was time the time was right and build this for your customer community >> having a huge uh expertise in uh data science, machine learning and artificial intelligence. Bear decided that it would be beneficial for uh the company itself for uh the bear ecosystem and also for the open market to build uh a supercomput which from the first day will be fully available through the public cloud and now we've just commissioned our second supercomput. So we now have two Chrisphari and Chrisari Neo supercomputers. When you started this, was the intent to provide supercomputingcapabilities to financial industry users or were you looking to expand to a broader set of commercial concerns in your region? >> We are looking to the whole market. It's not just be its ecosystem but also uh all the types of uh commercial organizations from big to small also some government agencies, universities. That's was our goal uh our goal from day one and uh our main vector of developing this was to bring the uh supercomputing power closer to the civilian analysts as we call them not uh just some hardcore machine learning or data scientists. >> I think a lot of us traditionally think of supercomputing we think of universities and research. you think of government organizations, maybe energy uh exploration, but for you to bring this to a broad-based business community is very exciting and it seems to open a lot of new doors and opportunities for innovation.>> We're having a number of uh health care providers. For example, during the uh start of the COVID 19 outbreak, we've implemented a free service which is based on machine learning and Christopher supercomputer which can analyze this lung scan uh and provide a second opinion for the patients or for the doctor based on the artificial intelligence and it was uh and is very popular. That's fantastic. Kind of a public service in essence to give that capability back to uh to health care tocome up with better solutions. That'sgreat. How did you come to include uh the NetApp and Invidia partnership in the ecosystem? Clearly in building this with the compute and storage and networking components and then all the layered software, you had to pick from some pretty capable supplier partners. Howdid NetApp and Nvidia fit into the mix? [clears throat] [clears throat] [clears throat] >> When uh designing and uh deciding on which way to go, we have two major options. One to uh create uh the supercomputer from scratch from some uh small scale components which may have helped us to save a bit in terms of the financial front and the other option is to uh utilize the existing and proven architecture based on Nvidia and NetApp. And we've decided to go this way which proven to be uh the best solution for our case because uh when building the uh first supercomput from the idea to the uh well to cutting the ribbon uh we've managed to achieve 6 months and uh then when deciding to uh build uh on how to build our second supercomput Chrisino we've stickked to the same approach and managed to do all that and 4 and a half months. So uh the time to market which is critical for our case was one of the uh key reasons to decide towards the proven uh Nvidia net app architecture. Most of us think about clouds and certainly large commercial clouds Amazon, Microsoft Azure, Google. Where does spare cloud fit in this larger ecosystem of services in your mind? >> Uh when transitioning from the uh own environment, own data center or server room, it's easier to go in small steps to transfer the workload from their own hardware, their own data centers into the cloud. Uh literally not changing anything in the software code. So that's the first step. Then as a second step we provide the huge variety of the platform services which includes uh managed databases, managed load balancers and uh the web servers as well as uh a number of security uh managed services to either build the new applications from scratch or to uh adapt own old legacy applications into the cloud native. environment. On top of all that uh we have the AI as a service uh capability which can be used with both infrastructure services and platform services or even can be used alone to provide the friendly and uh easy to use interface for the analysts of all types to perform the uh machine learning models training and inference uh either on the data stored in our cloud or uh from the other cloud or from on prem installations.>> At the base of what you're talking about with spare cloud is really an entire enterprise modernization capabilities and do you find that a lot of companies are in this hybrid cloud world where they're partially onrem and going to the cloud. It sounds like you're offering a lot of services to help those customers make a graceful u migration to cloud and cloud computing. Obviously technology is not the only thing holding uh the organizations from the from making that step to the clouds and then we step in with our uh additional consultancy services which we are able to provide to our customers to ensure that they find the robust, safe and secure way to uh transfer their data to the cloud. How do you enable these high-end data scientists and make it easy for them to embrace theseplatforms that you're offering? >> Obviously, we can provide the Jupiter notebooks which are uh well pretty much uh the way the hardcore data scientists prefer to work but the core functionality is uh moving towards the uh graphical interface. Also we have the feature which is called AutoML which contains no programming at all and the uh potential user can do all the modeling without writing a single line of code. And our main goal is to bring uh data science, artificial intelligence and machine learning closer to the people to make it easier for them to work on. Sounds like you'rereally improving that time to value proposition for data scientists and data modelers. As um certainly our NetApp clients started moving more aggressively to the cloud, one of their biggest concerns was security or data privacy. What can you tell our viewers about how these platforms are built to be secure private especially when you're talking about large quantities of data like this? So indeed the uh information uh security is the cornerstone ofall we do here because the safety of the data and the reputation we have uh cannot allow us to have a single failure in that field. We are coming from the financial ecosystem. This uh compliance and uh information security as a part of our DNA. I'm sitting here envisioning some of the really interesting applications of Christopher and thesupercomputing environment you've built, but give us some examples of some of the more interesting work you've seen done thus far. I think the most interesting uh thing we've just recently got uh trained on our uh Christophari supercomputers is the uh GPT3 which is the most advanced uh natural language processing uh algorithm and model uh which is based on the uh the full set of the Russian language literature and training materials. We were the first to introduce the generative AI model which can generate the images based on the uh words descriptions. >> We talked earlier about the NVIDIA NetApp uh components and partnership. How do you see partnerships in this space evolving? There is a specific entity which is called the artificial intelligence institute within the sperica system which uh provides uh the services for the universities as grants the uh medical application which uh provides the analysis of the city scans uh for COVID 19 patients and uh it is able to provide the second opinion for the doctors or for the patients from their financial sectors when we are providing the capabilities to uh build the financial models for the different markets. Some companies from the oil and gas sector who are building some geological related uh models. The huge application is the natural language processing for both uh voice recognition and text generation which is uh being practically used in the different types of the voice assistants within Barerica system uh and outside of it. Theodore, we'vetalked a little bit about AI and ML, but I'd be very interested in hearing what you see as trends or trending uh inthe customer space today in terms of those technologies. >> ML and AI is leading the transition, but all of that is uh just nothing without any simpler services. But the core difference that I notice for the Russian market for example is that the uh products which are based on the platform services are growing much faster than the infrastructure services. >> Can you tell us a little bit about how you build an environment like this to generate such confidence in 7x 24x 365 capability?>> I think it all depends on people and we have a great team. We are trying to attract the best talent but obviously that's not just the single point. Uh we have the quite uh sophisticated infrastructure build. We spread our uh services and workload through the different data centers connected uh with our own private uh dense wave uh multiplan network. Coming out of this pandemic, how are you attracting and retaining top talent to build such visionary solutions for your customers? >> Well, uh that's uh a cornerstone of our success indeed and we put a lot of effort in bringing in and retaining the talent. That's indeed a challenge with after the pandemics when everybody is moving around and it's easier to work for migrization abroad sitting from anywhere in the world. Our key stakes here are based on building the talent within. So we are trying to attract the uh young talent uh which maybe are lacking some technical skills but are eager to learn, eager to be on the cutting edge of the technology. >> You probably can continue to attract great engineers and great architects because quite frankly uh it sounds like a pretty neat place to work. I really hope so because uh we are really doing the most challenging and interesting tasks in IT I think in the whole country being a cloud provider and personally having the expertise from the our other side of the fence to encourage every organization to come and try some cloud services because it's really the enabler and we and when we are speaking about time to market it's hard to compete with. >> That's a great finishing thought. I really appreciate that. Theodore Proarov, thank you so much for joining CIO Central, for joining the show today. It's really been a fantastic conversation.>> Thank you, Bill. It's been a pleasure to talk to you. >> So, that's it today for CIO Central. We all hope you have a wonderful day and remember to stay data driven.