Kevin Mak is the co-founder of Ironfly Technologies, a company established in 2013 specialising in data aggregation, analytics, and trading systems for buy-side firms.
Previously Kevin was an options trader for a London based market making firm, and subsequently the senior trader at an Asia focused long/short hedge fund trading global equities.
Kevin holds a Bachelor of Computer Engineering, a Master of Biomedical Engineering, and a Master of Finance.
Who are you?
I am Kevin Mak, the CEO of Ironfly Technologies.
What does your firm do?
We started about seven years ago. When I was working at a Hedge Fund in Hong Kong we used an EMS called Portware, now bought by FactSet.
My background was not actually financial, I was an engineer, I grew up playing computer games. And then as a trader I would spend 12 hours a day staring at a screen and I just felt that the system needed to be better than spreadsheets. That was the genesis, the why, of Ironfly Technologies.
I get it, it’s difficult, there are lots of moving parts, lots of stuff under the hood. You are effectively mapping software to the messiness of the real world, different exchanges and it’s tricky. But, nonetheless so much of that effort goes into “that order has been routed, great now let’s find some weird exotic thing to customise for this customer”. I understand why that naturally happens but we try to keep focused on a philosophy of decent design - and by design I don’t mean “it looks pretty”. That aesthetic part is really a by-product from things being in the right place, the important things being big, the unimportant things being small and that sort of thing. But in general, we have much more of a design focussed way of building out that workflow and the system overall.
A lot of what we build ends up being very visual – “a picture is worth a thousand words”. How do you make spreadsheets better? That’s where all of the real-time visualisation of data came in. Ultimately this is about: “you have a lot of information, you have a bunch of work to do, how do you make that work for you?” That’s all about how you understand that information before you go and do something with it. And that’s where it all came from really, a bunch of nerds sitting around saying this should be better!
You were at the sharp end of using trading technology from a bunch of different vendors and then decided to do it yourself. If you could go back to a younger version of yourself, what would you say to him?
That’s a good question! There are two things I would say. The first thing is – don’t do it! [laughs]. The second is, if you are dumb enough to go and do it anyway, then don’t stop. Whatever you do, don’t stop. You never know when things are going to work, when you will get that big client. It’s just that you don’t know - sometimes it’s about just hanging in there. Especially for us vendors, the types of people we work for, we have clients that have seen the product, not signed up initially and yet have now come back to us. They’ll will say “that we spoke to you when you were just three years in and we didn’t know if you would hang around… but you are seven years in now, so ok, you guys didn’t die, congratulations for surviving, we can now talk to you and let’s do business.”
It’s just one of those things about this industry, flow begets flow, that’s the mentality.
Technology-wise we would not change a lot, our stack is pretty good and we have gone in the right direction. We would probably have made a couple of changes with how we handle market data. At the beginning we were a little overly optimistic about how simple market data could and should be – which it’s not [laughs]. So everything from Reuters feed to Bloomberg SAPI is a complete mess. Data should just come and it should just work, but no, it doesn’t and it’s hideously expensive.
Otherwise, going down the rabbit hole… It took me a little while to really identify what it is that we are really doing. You can say trading systems and that’s fine, but at least from an overall philosophical perspective, we are there as decision support. So how do we do that? How do we support you to make more decisions, faster decisions, better decisions? We don’t advise, we don’t say “you should buy this and sell that”, that’s what you have analysts for. With the information you already have, how can we communicate that back to you in the most effective, quickest way? That’s one of the things that’s very hard to articulate about our product. Here’s where we are different, that we focus on this as a core part of our DNA. We train our developers to think about this and applies to how we design the actual product. Yes we support FIX and algos, but once they are there in the system, we go back always to “how do people actually use this?”. It’s more about the design focus and if I was to change anything, I would say to myself to try and articulate this design thinking message earlier on. Rather than just saying we do trading systems as it’s harder to differentiate yourself in that respect
What do you do for the firm?
Answering that depends what building I am in - if it’s likely that the person who asked is going to understand capital markets then it’s a different answer to if I am at a dinner party and I need to be more interesting to them than saying I am a dentist.
I usually say “we take financial data and make it into real-time pretty pictures”, that’s the glib one-liner. It then leads to the next conversation – why do people need this in real-time, why do people need pretty pictures. And then you can jump into the why do we do that, that humans understand data in certain ways and spreadsheets are often not the right way. Even when you use spreadsheets you try and line them up, gets the fonts matching, get the numbers lined up so you can see from a size perspective how big a number is relative to the peer group before you even read the number. We explicitly do that, we generate that picture on-the-fly as the portfolio changes, as your orders change.
How to make that a fun, dynamic, interesting conversation, depending on who that person is, that’s not always easy!
Day-to-day what do I do for the firm? When we started we were all very technical, so for the first few years I was in amongst the code. These days I do that a lot less. Up until about three months ago I would have said it was about half of my time in the code. Now it’s probably down to about one-third. We are building out the engineering team, it’s bigger now so that really helps with transitioning that work off my plate.
The next phase will be, apart from all of the general management things, to focus more on becoming a lead generator myself.
How do you hire?
When we hire developers we focus on the underlying technical ability and skillset rather than ‘oh, you’re from finance, great’. Whenever we recruit we’ll receive CVs and recruiters may say “this person has ten years of Java experience in a bank”. That’s great, but we run a web application, so we don’t need that experience, we have maybe 5% of our codebase in Java. So currently we have to teach a lot of institutional finance to the developers we hire. That’s fine, you pick your poison and this allows us to do more fancier things on the technology side.
What made you join your current firm?
My education was never geared to finance, I studied computer and biomedical engineering. I was more of a nerd than a finance guy. When I left University I found all of my old high school friends were telling me that they had joined big investment banks on $100,000 graduate salaries. I thought, well ok, you’re smarter than me but not two and a half times smarter than me, that’s not so cool [laughs]. So I went off and joined a market-maker, the only other people in finance that will not exclusively take law graduates or business people. Not knowing anything about finance helped in terms of having an idealism in the expectation of what systems should look like. But no, it’s just spreadsheets and more spreadsheets!
The market making firm gave me a desk with eight monitors all with spreadsheets. If one of the cells was highlighted in yellow then maybe you want to click it! And that’s what you do all day. You would think, given the amount of money going the systems, they would spend a little money making it a little less error prone, more usable - but that never happened.
I relocated to Hong Kong and joined a Hedge Fund… same problem! At some stage I realised, I grumble about this three times a day so either put up or shut up. And meanwhile - not by ability but by attrition - I found myself moving up the ladder at the fund to the point where I had to sign off for things. And then I hired a couple of traders and needed to sign off for two new Portware logins – for HOW much? Is this in Hong Kong or US dollars? This is in US dollars? And they charge the broker side as well?
Finally, I understood just how much vendors were charging for the system. I thought, I can calculate expected value. If I stay at this firm doing this job I can make X amount of money, which is pretty good, but if I go and do a different thing, set up a company and solve this problem by building a product. Well then I get to build a product, which I haven’t done before, and also the financial numbers back up that strategy. Even if there’s a one in four chance of making it work it’s a no-brainer. It would be irresponsible not to do it.
When we started, fintech was not yet a commonly used term, so being a first-time entrepreneur I didn’t really know what to do. When you have been on the buy-side and you are leaving a firm, your sell-side brokers hear about it. So they all asked, “which fund are you going to?” They all assumed it would be a competitor but I said no, I am going to do this. And they all said “hey that sounds pretty cool, we hear people complaining about their systems all the time too.” Some said – “do you need money?” I thought “wait people just give you money??" This could work!
We raised a little bit of money, from people within the industry, and got started. It wasn’t a big raise but it was enough to keep a small team going for a year-and-a-half. And that’s all we needed.
What do you see going on in the world right now where people are gaining real commercial advantage from the product/services of your firm?
I could point to us having better technology and workflow but people reading may fall asleep. Right now people get the most benefit from us because I happen to have done what they do before [traded] and I have spent a lot of time training up the people who are supporting our clients. We can read between the lines when a trader, portfolio manager or analyst asks a question about the software or says “we need a thing”. One of the issues we had when we were talking to Portware - and everyone has the same problems with vendors - is that typically your account manager or front line person, if they have not done it before, they can’t be left with just “hey there is something weird with the system, I’ll leave you to look into it”. This lets us deal with it faster, rather than saying that “we will see what it is and investigate, or that it’s a weird FIX message, let’s check with the FIX network people” and then it’s three hours before you get an actual answer.
I am hesitant to say this, as while we scale it naturally becomes more like a big firm and you have the processes, controls and the systems and so we might lose some of that direct expertise. But, at this stage, if I am being very genuine, that’s the actual answer – if a client takes to us, they don’t need to talk to us a lot. We’ll listen, get it and get it done for them.
Having said that – the product is really good too!
Blockchain / Distributed ledger technology – in use with your firm? Hype or real?
It’s not something we work with directly, but at the moment we are working on a project that touches on that field. We are in the early stages of talking to some exchanges about some work in that area.
The best way I can put this is that blockchain is a fantastic solution that were all trying to find a problem for.
It can definitely be there, but in many implementations it’s used more as a database. In which case, why? Just use a database!
For me the key will be in primary markets - primary market listings - to start being put on smart contracts. Right now we have some smart contracts as proxies for swaps, which can synthetically mirror existing stocks. But when primary markets start using these things in earnest that will be very cool, since then we will have a whole new class of open assets that can be traded anywhere in anyway. You won’t have to pay millions to list on an exchange. You will still have those guys around as some larger listings will want the prestige of listing on a traditional exchange such as NYSE or Nasdaq. But someone who wants to raise money to go from one pizza shop to ten pizza shops, that will be a whole new thing, traded on a very tiny exchange, there’s no physical infrastructure needed. I definitely see that happening, I hope in the next few years.
I feel very strongly about this as it breaks not a quite a monopoly but removes excesses and inefficiencies with the traditional exchanges within a heavily regulated system. That system needs to stay, but there is scope for smaller players to do smaller things around it. I look at this very keenly as ultimately we are asset agnostic, yes we do a lot of exchange listed products as that’s where the demand is right now. But once these new assets start coming in, I fully expect we will start plugging them into our platform. So the same interface for trading your BHP, CBA, APPL, your crypto all in once centralised system on your desktop.
Cloud - in use with your firm? Hype or real?
Very much real. We deploy using Docker, so we run containerised instances of the code for each of our clients. , Most of the time we run and maintain it but the client can also run it themselves. You can run it on AWS, IBM, Azure, whatever cloud service you want. Or we can deploy to a client’s own datacentre. For some of our clients with their own datacentres we just deploy to their datacentre and we can manage it remotely.
We offer a browser based front end and a fat-client, the application is really just a repackaging of the chromium engine. It’s a little like OpenFin in that regard.
Where do you see technology impacting in the next five years on:
You
Your firm
Wider market
Me personally, I never buy the latest phone. I like to buy the new version of last year’s generation. The reason why is that I have a paranoid mind – have you ever heard the saying “two is one and one is none”? I like to have backups of everything, I always have two phones. I don’t carry them both with me but I have one from a year and one from the previous year.
Technology wise I love seeing new stuff come to market, not because I will rush out to buy it, but because that will be the normal thing in two years’ time. That’s why I love seeing more computing power, whether on mobile or desktop or wherever as that ultimately allows us to do more as that technology will trickle down in two years’ time to the mass market. And for us as software guys, more compute means we can do more cool things. A lot of the time we are hampered by poor internet connections or bad laptops that don’t last very long and all of that is going to change. I am always excited by that.
Things become better when tools that a very specific industry uses become accessible to everyone. A lot of the time the way I view it is that a lot of financial management products may bamboozle a user. I say that alternative assets is anything my Grandmother does not understand. A short-sell? She doesn’t understand that, so it’s an alternative asset. One way is to educate people, another way is to make it more safe, simple and accessible. And that is what we do for the institutional side of things. And there’s no reason why, as our tools become better, people cannot do their own portfolio rebalancing or understand the risks if they are, for example, highly exposed to US Technology stocks. These are analysis tools that we are giving to professional traders now but there’s no reason why in some guise, whether we do it or someone else does it, these tools should really trickle down to everybody else.
And that is the sort of stuff that I think is more game changing.
Go crazy – make some wild predictions about machine learning, AI, DLT, cloud, whatever
A couple of years back we experimented with a neural network. Very simple, just tensor flow, a logistical regression neural network that we used to experiment with a data set from a large asset manager in Japan where they were doing a couple of thousand trades per day. We posed the question – is there any discernible information that I can pull out from this data set to suggest that some brokers are better than other brokers when it comes to trading Japanese stocks? Short answer (spoiler alert) yes, there is. We got to predictive accuracy at a high 60% - much better than chance, which is great. But then – now what? And what we found is that the effort taken on the analysis was only the beginning, and that the productisation to assist in decision support for a particular order would be the real value – so if you were trading a specific stock then the analysis might suggest to use a particular variety of a VWAP algo in order to get an optimal outcome.
So, the technology is there, but there’s a lot of hype and in most cases it’s a lot more work than people think it will be.
The second paragraph (biography) was supplied by Ironfly Technologies as was the picture of Kevin Mak. This interview was conducted remotely over a Zoom call and transcribed by Alignment Systems. No payment was asked for or received either way.
Comments
Post a Comment