Dr. Lee Braine has spent the past seven years working across Barclays' wealth management, markets, and corporate and investment banking divisions but his job couldn't be further from that of your typical City of London broker or trader.
A managing director in the bank's chief technology office, London-based Braine is responsible for research and engineering across corporate, investment, and retail banking. He has a special focus on distributed ledger technology, like blockchain, as well as quantum computing.
Insider sat down for a virtual chat with Braine who was named one of Business Insider's 100 Transformers to discuss what he's working on and where the industry is going next.
Transcript has been edited for clarity and length:
Insider: You're not an average banker. You're a computer scientist by training. Can you tell me a bit more about your background?
Braine: I have a PhD from University College London in computer sciencethe particular topic was object-oriented functional programming. I used that research knowledge in banking, and that included, for example, working with financial market infrastructures when I was in my twenties to produce new architectures and new optimization algorithms. In this case, it was for securities settlement. After that, I spent quite a few years working in technology management.
Within Barclays, for these last 7 years, I've been working on technology innovation. The typical thing I've been working on is responsibility for advanced technologies that Barclays needs to be up to speed on. We work closely with a variety of stakeholders, not just [technology] vendors, but also very closely with official institutions, including central banks, regulators, and the government on the potential of these new technologies, and the risks and any issues that may lie with them.
Insider: Let's talk about distributed ledgersyou lead Barclays' efforts there. Why is Barclays interested in distributed ledgers like the blockchain?
Braine: Interest was initially sparked about 5-6 years ago when we were looking at bitcoin from a technology perspective. That means not as an investable asset, but at other interesting, novel technologies underlying bitcoin that could be repurposed in more traditional financial services. There are several features of bitcoin that inspire a different way of working: at the lowest level, there may be things such as consensus algorithms, hashing technique, the chain of blocksall of those types of low-level technical things that everybody learned about in the last few years from blockchain. But higher up, there are new ways of working, almost new market models that get inspired by cryptocurrency.
For example, currently, financial market infrastructures are centralized financial institutions, and their technology is centralizedthey've got centralized databases and centralized processing. The decentralized nature of something like bitcoin has inspired people. Could we have a different model of the market? Could we imagine decentralizing, not just the technology, but also some of the rights and obligations of participating in such a network? So to make that abstract idea a bit more concrete: imagine if you've got a clearing house, and currently we send all our trades to the clearing house, it performs the processing and sends us back the result. Imagine if, instead, each of the participants formed a network, they operated peer-to-peer, and that peer-to-peer model then gets translated down into the technical solution. So that's a different way of workingyou can call that a distributed financial market infrastructure.
It's a big infrastructure change to the marketso why bother? What we see is quite a few potential benefits. These include radical simplification and rationalization. Another thread is you're able to speed up settlement times.
Insider: Tell me about Utility Settlement Coin and the Fnality investment.
Braine: The consortium was originally called Utility Settlement Coin, and then, about 2 years ago, a group of financial institutions so 14 banks and one exchange strategically invested to create the new entity, which was Fnality International. They're building a new payment system, and this is going to offer peer-to-peer settlements using an underlying blockchain platform. The money that moves on it will be one-to-one backed by funds that have been pre-deposited at a central bank, so it's effectively a pre-funding model. It allows a number of benefits in terms of settlement.
For example, you could continue operating outside of the window when the real time gross settlementRTGSis closed at the central bank. You could, for example, connect to other tokenized assets to allow atomic swap between them. If you had Fnality representing the payment leg on a payment blockchain, you could imagine a security leg on a security blockchain and the two of them could do instant settlement with the appropriate interconnect between the two. A key point here is that the money being backed by funds at the central bank means that there's lower risks associated with such payments.
Insider: You also work with the International Swaps and Derivatives Association (ISDA), right?
Braine: Yes. One of the things we've been progressing for a few years relates to a new standard for data and processing, and it's called the ISDA Common Domain Model. This model effectively provides a standard industry representation for events in the lifecycle of a trade. Currently, each institution builds their own solutions, so effectively, there's variation in how you code itsome may code in Java and others may code in C++, so different programming languages. They may store the data in different types of databases, and they may enrich the data with extra fields. So you've got variation there. Then, over time, each institution must manage and maintain its data stores. So across the industry, the same high-level functionality is implemented slightly differently on slightly different data sets. And each time there's a lifecycle event, they all need to sync up and reconcile to make sure that, yes, what's been affected in terms of an event, the before and after, is consistent.
That's incredibly inefficient as a solution. Imagine we had a browser for the internet, and each bank built their own browser, right? Of course we don't do that. We have a common browser, Chrome or Internet Explorer, we download it, we use it. So that same philosophy is being applied here. A distributed ledger de facto defines the common data structure that you all must use. And smart contract technology is a common process that they must all follow.
You then start getting the opportunity to transform the industry, and all the participants. And those opportunities don't come up very often. So I think we're living in interesting times where this technology is just reaching the right degree of maturity, and there's also appetite from the market participants to reduce costs.
Insider: Ok, tell me more about smart contracts, which I know you also research.
Braine: There are many, many business processes that could benefit from the rigor and standardization that smart contracts would bring. To give one example, interest rate swaps. So a few years ago, about 4 years ago, my team prototyped an interest rate swap from end to end. Complete end-to-end processing naturally fits with the idea of a smart contract, meaning the data that you construct at the beginning just flows throughyou don't transform it, you don't switch it into completely different systems.
The way I like to view it is, smart means automatable, and contract means enforceable. Other good use-cases include trade finance, loans, bonds, and syndicated loans. It's easy to identify 101 use cases for smart contracts; the challenge is identifying viable business cases where the industry can move together in concert, given that these are consortium plays, so you need your peers to be similarly motivated at the same time to grasp at the same propositions.
Insider: What sort of work are you doing in quantum computing?
Braine: Barclays started exploring quantum computing back in summer 2017. We did that by partnering with IBM. We set up a joint development project, and our goal initially was to learn more about quantum computing. It's a phenomenally complex topic, where even those that have quantitative research backgrounds find it challenging to understand the details.
We decided for our first proof of concept that we would look at a settlement optimization problem. This is a particular challenge where a market infrastructure looks to optimize the settlement of a batch of securities transactions. A typical batch may have 50,000 transactions, you've got many potential combinations that you could settle, and you need to work out what is the best combination. It's a problem that you typically can't solve perfectly, so you often run an optimization algorithm for long enough in order to solve it well enough, and then you repeat the batch later.
We were inspired by [the question]: could a quantum algorithm on a quantum computer solve that problem perfectly, or perhaps better than the classical ones? We looked at candidate quantum algorithms, we worked with IBM to implement an algorithm, we constructed candidate scenarios to run through test data, and we got the results. The key takeaway is that, for the first time, an algorithm has been run for settling securities transactions on a quantum computer. Obviously, it's only just test data and very small scale, so it's more of a proof of concept, but we've demonstrated that the proof of concept works.
In terms of next steps, we're currently exploring quantum machine learning. How many more buzzwords could you get into one conversation, right? We've run our first experiment comparing quantum and classical versions, and in the next couple of months, we'll be looking to publicly release our initial findings.
Insider: In real terms, what benefits might quantum computing bring to Barclays? And when?
Braine: We need to extrapolate for when we think the hardware will be sufficiently mature to be able to run real-world use cases. For perspective, we think that will be in the range of 4-8 years from now.
In terms of the type of benefits, it's almost like adding a special maths co-processor, and it's able to perform a number of functionsit's able to perform an optimization process faster than a traditional classical computer, or it's able to perform the process and get a higher-quality result. So this could be optimizing which assets you put in a portfolio, or running a number of Monte Carlo optimizations as part of a risk model. These types of things often require huge compute resources.
And that's why we're exploring this for researchnot because we think it could be perfectly used in the next year or two, but because we're learning, building a foundation. I would almost call it quantum awareness, where we're raising our awareness so that we could leverage it when the powerful machines come along in a few years' time that we could use for real world use cases.
Insider: Where is financial technology going next? How does that fit with traditional banking?
Braine: There are a number of key themes, one being machine learning and artificial intelligence. So the application of that technology, we've seen it deployed with fantastic effect, whether that's search or shopping or similar. There's great opportunity for those technologies to also be applied within financial services, particularly to further improve the customer experience.
Other key technologies include cloud computing.
Insider: Are you worried about disruption from tech startups?
Disruption is at the heart of my day job. We're often looking to see what technologies have potential for disruption, and to see how we could leverage or partner with third parties that have such potentially disruptive technologiesand also to understand the risks and potential issues that are associated with themso that we're able to have a sensible position in order to be able to advise the business.
Often if we're going to pick certain technologies, it may well be the case that what's viewed as a potential disruptor to Barclays could also be viewed as a potential partnership opportunity in terms of optimizing and improving some of our own internal processes.
Read more here:
Meet the Barclays MD working to transform finance through distributed ledgers and quantum computing - Business Insider
Read More..