Whatever you think of Michael Lewis’s book Flash Boys, there’s no denying its significant impact on the overall perception of the high-frequency trading sector. Before the book, even many markets professionals understood little about the strategies and techniques employed by the arcane fund fortresses of rocket scientists who execute a huge number of trades in nanoseconds. Subsequent to the book’s release, the HFT world has needed a fresh face, an ambassador that can show that high-frequency strategies are not incompatible with fairness and basic transparency. That’s why The FR’s Gregg Schoenberg was delighted to connect with Christina Qi. Over the course of a few conversations, Qi, the CEO and co-founder of Domeyard, painted a compelling picture of her company, which she started alongside Luca Lin and Jonathan Wang. Today, with big-name backers including Gary Bergstrom (Acadian Asset Management), Howard Morgan (Renaissance Technologies and First Round Capital) and, more recently, Softbank, Domeyard’s stature exceeds its AUM or headcount. However, it has captured the interest of the greater financial community for its distinct managerial philosophy, embrace of AI and its positioning as a next-generation hedge fund.
The Financial Revolutionist: Christina, it’s great to connect with you again. I’d like to kick things off by asking what made you and your two co-founders believe that you could build a successful high-frequency trading firm without the financial firepower that many of your peers possess?
Christina Qi: Thanks, Gregg. We had interned for financial firms and trading desks and had noticed how slow they were to adopt technology. We had also been trading on our own for some time and naturally progressed from there. We weren’t driven by a passion to compete against other people, though. We just wanted to combine our passion for technology and the markets and see if we could build a venture out of that.
FR: Plus, you were straight out of college and could take a risk, right?
CQ: Yes; it’s less risky than it seems, because we realized we’ll always be able to find work at a large bank or tech firm, which wouldn’t be so bad.
FR: They pay well, but it’s not the same thing as running your own show, and from what I hear, things have worked out very well for Domeyard. Plus, you now have Masa Son as an investor. I’ve heard meeting him can be a nerve-racking experience. Have you met since Softbank acquired the 15.92% secondary stake in Domeyard’s management company?
CQ: We’ve spoken with his team. It’s pretty crazy to be a part of a portfolio that includes WeWork and Uber. I also think we’re the only HFT firm in their portfolio.
FR: That’s probably true. Moving on to your performance, I know that you need to be careful about discussing your track record, but I’ve heard from various sources that your performance has enabled you to charge quite a bit for access to your fund.
CQ: Yes; we cannot disclose everything, because we’re regulated and subject to certain rules on non-solicitation. This is hard for me, because as a Millennial, I’m used to sharing everything. In terms of access, it’s because many of our LPs could have taken up the entire capacity of our fund given our size. The fact that we’re capacity-constrained can be both a good and bad thing, but it’s enabled us to choose our investors and only work with those people who share our long-term vision, which is to build the hedge fund of the future.
FR: It sounds like you’re a Millennial who wants to keep her very cool job. Can you at least speak to your average daily trading volume?
CQ: I don’t know the average because it varies; however, we were getting up to about $1.5 billion in volume per day when I last checked.
FR: As I’m sure you’re aware, the quant sector has had a rough go of it lately. How are you feeling these days?
CQ: It’s not a surprise given the shift in volatility. But to your question, we’re relatively happy, although I don’t want to compare ourselves to anyone else, because that’s not the way we measure things.
FR: On that note, in poking around your web site, it’s pretty clear to me that you’re trying to differentiate Domeyard from other HFTs. Do you think the industry deserves the bad rap it’s gotten?
CQ: Oh yeah, definitely. I think many people have the wrong impression of our industry because most HFTs are overly secretive. We’re not afraid to be more open and come right out and say that we’re an HFT. But we’re also very clear that everything we’ve created has been done very ethically and fairly — none of that front-running or quote-spoofing that some imagine. Basically, we’re just a different fund than most everyone else.
FR: Partially through the use of AI and machine learning, right? Could you shed some light on what you mean?
CQ: Actually, many HFTs don’t use any machine learning in their alpha generation processes. The typical strategies are very simple and do not apply much statistical modeling. Most high-frequency traders who depend on structural pricing relationships have little use for ML-based predictions, but we think there’s a lot to gain in terms of strategy capacity by incorporating ML.
FR: Meaning that ML strategies are tough to implement, right?
CQ: Yes. I think high-frequency strategies that employ machine learning entail more than double the effort to implement, because they not only need to be fast, but also require complicated modeling. The models themselves can be customized to be very fast, which is what we try to specialize in.
FR: So how are you using it, then?
CQ: We use machine learning to dynamically calibrate our signals throughout the day, and to determine which signals to deploy and which ones work best together. Besides that, we — and a few other firms are doing this now too — use natural language processing techniques for textual analysis.
FR: So processing unstructured data and converting it to actionable information.
CQ: Yes, converting text to legitimate signals that can be used.
FR: Do you subscribe to any particular AI discipline?
CQ: We employ all kinds of disciplines. There’s a sizeable use of deep learning, but I don’t want to overstate our use of it either.
FR: Do you think there’s going to be a paradigm shift in HFT in which the real advantage will move away from milliseconds and toward those who have an information processing edge?
CQ: Definitely, yes. We would word it by saying that there used to be this race to zero. But these days, anyone trading from a dorm room can pay a vendor to get access to incredibly fast speeds. Maybe not the fastest, but fast enough.
FR: Some firms, though, still play the latency advantage game.
CQ: That’s not our focus, but yes, there are those firms where maybe every two years, a new piece of hardware will come out, and they’ll spend a few hundred grand to acquire it so as to be tens of nanoseconds or a few hundred nanoseconds faster. That’s great, but there are other ways to make up for the disadvantage of a few hundred nanoseconds.
FR: I take it some of the problems with the speed game are on the exchange side.
CQ: That’s part of it, yes. Even if you place an order a few dozen nanoseconds faster than your competitor, the exchange will take forever to process it relative to how fast you are, and there’s a high probability that your competitor’s order will still arrive before yours because of the large variance. So really, the speed advantage goes away.
FR: As we look to a future for HFT — where information processing is the key — do you think that some entity out there — maybe it’s Domeyard or maybe it’s someone else — could gain some type of winner-take-all edge? A Google position of sorts?
CQ: That’s a good question. First, I’d say that we do think that the industry is maturing and that we’ll start to see who the real winners are. But more broadly, this isn’t a winner-take-all market. People always ask us, ‘How can we compete against the Citadels and Virtus out there?’ But even our largest competitors find it challenging to sustain over 10% share of market volume, and we don’t see that changing.
FR: You probably get asked this question a lot, but has Domeyard looked into crypto?
CQ: Any kind of shop involved in technology trading should be looking into it on some level.
CQ: Because the underlying technology is here to stay. For us, we’re probably not going to be trading arb on the exchanges or trading crypto over the next few years, but we’re collecting data.
FR: The reason I ask is that because your LPs are mostly family offices and ultra-high-net-worth individuals, you have greater agility than a peer with lots of institutional backers.
CQ: Well, it’s true that we wouldn’t have to start a new fund to trade crypto, but we’re very cautious about it.
FR: Why is that?
CQ: Several of our investors have been asking us about this, too. Look, 1000% annualized returns sound amazing, but we tell our LPs that it’s extremely volatile and prone to price manipulation, the capacity is low, and there are a lot of questions about how this market will be regulated that are still unanswered.
FR: Well, caution is always good, particularly in a time of rampant speculation and volatility.
CQ: Yes, not to mention there’s an enormous amount of hype in crypto. The market has such a low barrier to entry that many people who want to start a crypto hedge fund have no real trading experience. I’m approached almost every day by someone who says, ‘I’m going to start a crypto fund. Can you be our board advisor?’ I’m quite skeptical that they’ll have sustainable edge like this and urge caution.
FR: Let’s turn to Domeyard’s longer-term goals. I’ve noted that even though you’re structured as a hedge fund, you also resemble a fintech start-up in some ways. Can you elaborate on that?
CQ: The start-up-like component is that we’ve built a lot of IP. The trading platform, risk management systems, and feed handlers were all built in-house, which provides a lot of value for our investors and for us. In fact, I don’t think we’ve outsourced anything related to the trading part of our company.
FR: Have you considered using those proprietary systems for uses beyond trading?
CQ: In the future, there are a lot of use cases for our data analysis. Maybe that will be in biotech, or in creating long-term hedge fund strategies, or in generalized fintech.
FR: I think you mentioned when we last spoke that you are licensing out your trading platform to at least one entity, right?
CQ: Well, we are in beta testing mode with a few fund managers that are licensing our technology for testing out some strategies that they have. The goal for us is to see if this can become a viable business model for us over the long term.
Role within the HFT community
FR: Let’s turn to your role. I don’t have to tell you that there aren’t a lot of Asian female CEOs of US hedge funds. Do you feel added pressure because your success could inspire a lot of other women?
CQ: I see it more as a responsibility than as a pressure, given the unique position I’m in within the high-frequency trading industry.
FR: You definitely don’t look like the typical HFT heavyweight.
CQ: It was something I didn’t notice until recently. We were looking for board member recommendations for our company, and all the names I received were men. So I reached out on Linkedin to obtain some suggestions of experienced women in the industry too, just to have a diverse candidate pool. People then wrote back to me saying, ‘Christina, I think you’re the most successful woman within the HFT industry.’ In quant trading, sure, there are some successful women, but in HFT, there aren’t many.
FR: Thus the responsibility you feel.
CQ: Yes, but my broader goal is to influence how this industry is perceived in the media. HFT still has a terrible reputation, which probably won’t be helped when the Flash Boys movie hits Netflix.
FR: No, probably not. You’ve taken special pains on your site to convey that Domeyard has a flat managerial structure.
CQ: We created a flat structure for a few reasons. One is that I’m a very young person — I’m 27 — and the average age of our people is considerably older. But we’ve been able to attract amazing people by saying to them that they will be treated as leaders themselves.
FR: Giving them a sense of ownership.
CQ: Yes. It’s a great privilege to have people twice your age say that they want to work with you and that they respect you. In fact, we now get 20,000 to 30,000 job applications per year.
FR: That’s a huge amount given the size of your firm.
CQ: Yes, but the organic interest in Domeyard gives us the chance to find people who can fit in well with our flat structure and who have the right skill set. We purposefully don’t hire any rockstar with a huge ego. That helps us keep the flat culture working: nice people who are humble, not siloed or competitive.
FR: When you meet with investor prospects, how do they react to your culture?
CQ: We’ve had at least a hundred investors visit our office in Boston. Most of them have been great, but we’ve also had investors treat us like kids, which isn’t such a bad thing, but it’s been used in a way to create a power dynamic that is unfair. I mean, we are very experienced at this point, especially because your learning curve is accelerated when you have to build everything from scratch as a start-up. But unlike a male my age who can just grow a beard to look older, I can’t do that.
FR: Well, hopefully you’ll be able to attract some women into your space.
CQ: Absolutely. One final thing I wanted to mention: We treat other HFT competitors very well, despite the competition and secrecy. I talk to a lot of CEOs who have been very cold to me. This is unwise, from our perspective.
FR: Why is that?
CQ: There’s a lot of consolidation going on in our space right now, and it just doesn’t make sense to go around being cold to people. You never know if you’re going to need each other’s help or work together down the road.
FR: So is it fair to say that creating positive karma is part of Domeyard’s alpha generation strategy?
CQ: Yes, that’s a perfect way to think about it.
FR: It’s been a pleasure speaking, Christina.
CQ: Thanks so much, Gregg.
This Interview has been edited for content, length and clarity.