Weldon Rice
AI-Powered ETFs: Machine Learning Picks Stocks
Weldon Rice heads up AI ETFs at QRAFT Technologies, a Seoul, Korea-based investment technology company founded in 2016. Weldon's path to finance was unconventional: he moved to Asia right after college, lived in Korea for over 12 years, and studied under quantitative hedge fund managers and Goldman Sachs's Korea operation. QRAFT received a $140 million investment from SoftBank, which accelerated the firm's expansion into new asset classes and jurisdictions. The majority of QRAFT's employees are researchers and developers who build AI models entirely in-house.
On this episode, recorded live at the Exchange ETF conference, Weldon talks with Brad about how QRAFT uses AI for stock selection (not as a thematic bet on AI companies), the difference between analytical and generative AI in portfolio management, and their two ETFs: QRFT (US Large Cap) and AMOM (Large Cap Momentum).
What QRAFT Actually Does with AI
Weldon starts with an important distinction: QRAFT is not an AI-thematic fund that buys AI companies. It's a fund that uses AI to pick the stocks in the portfolio. The firm positions itself as "quant 2.0," taking academically researched factors that have been around for decades (value, momentum, quality) and using machine learning and deep learning to build more reliable exposure to those factors with better risk-adjusted returns.
QRAFT's AI sits on two sides. The analytical side uses pattern recognition on structured numerical data: price, market data, financial ratios, and other quantitative inputs. This is the core of their stock selection engine. Traditional quant models rely on linear regressions, but QRAFT is building non-linear models because, as Weldon puts it, "we know that markets aren't linear." The deep learning side adds more complexity, with neural networks that can identify patterns across larger datasets and more variables than traditional machine learning approaches.
On the generative AI side (large language models for text and sentiment analysis), QRAFT has some capabilities but has been more cautious about deployment. They recently partnered with LGI Research specifically because of LGI's strength in large language models, launching a fund that incorporates sentiment data from news sources into the AI prediction model. Weldon emphasizes they don't add AI features just for marketing purposes. Every AI input has to earn its place by demonstrably improving the model's predictive accuracy.
QRFT and AMOM: The Two ETFs
QRFT is QRAFT's US large cap ETF. It uses the full AI engine to select stocks from the large cap universe, rebalancing on a regular cadence as the models update their predictions. The fund doesn't make predictions beyond about one month, which Weldon explains is a deliberate constraint. Their research found that prediction accuracy degrades rapidly beyond that horizon, so they keep the window tight and update frequently rather than making long-dated bets.
AMOM is the large cap momentum ETF. It uses QRAFT's AI to identify momentum opportunities within the large cap space. The key innovation here is that the AI can identify when momentum signals are likely to work and when they're likely to reverse, which is the classic problem with momentum strategies. Pure momentum gets crushed during momentum crashes (like Q4 2018 or the COVID rotation in late 2020), and QRAFT's model aims to reduce exposure before those reversals happen by reading early warning signals in the data that traditional momentum screens miss.
The SoftBank investment allowed QRAFT to expand across asset classes and geographic markets, extend the frequency range of their strategies from intraday to monthly, and build out more sophisticated capabilities like long-short strategies. Beyond ETFs, QRAFT operates as a B2B solution provider, building AI-powered model portfolios, robo-advisors, and portfolio construction tools for large financial institutions in Korea and increasingly in other markets.
Guardrails and Continuous Improvement
Weldon addresses the natural question about AI guardrails. The models are constantly updated with new research and data inputs, but there are constraints on what the AI can do. Position limits, sector concentration limits, and other risk management rules are coded in as hard constraints that the AI cannot override. The team of researchers continuously evaluates new data sources and model architectures, but any changes go through a rigorous validation process before being deployed in live portfolios. It's a balance between letting the AI learn and adapt while preventing it from making outsized bets that could blow up.
Key Takeaways
- QRAFT uses AI to pick stocks, not to invest in AI companies. Their models combine machine learning and deep learning to build non-linear factor exposure, which they call "quant 2.0."
- The firm received $140 million from SoftBank, growing from 40 to significantly more employees, with the majority being researchers and developers who build all AI models in-house.
- Predictions are capped at roughly one month because accuracy degrades rapidly beyond that horizon. The models update frequently rather than making long-dated bets.
- AMOM uses AI to identify when momentum signals are likely to work or reverse, aiming to reduce exposure before momentum crashes that destroy pure momentum strategies.
- QRAFT also operates as a B2B provider, building AI-powered portfolio tools for large Korean financial institutions, making ETFs just one part of their broader technology business.
Listen to the full conversation on Spotify, Apple Podcasts, or YouTube.
Full Transcript
4,237 wordsMachine transcribed from Brad Roth's conversation with Weldon Rice, with speakers identified automatically. Timestamps link to that moment on YouTube. Lightly cleaned, otherwise unedited.
Welcome to Behind the Ticker. I'm Brad Roth, Chief Investment Officer of Thor Financial Technologies and Portfolio Manager of THLV, the Thor Low Volatility ETF. Behind the Ticker uncovers the inner workings of the ETF industry. We will interview portfolio managers and ETF service providers to dive deep into their work lives and their businesses. We will learn the inner workings of their strategies and what drives them as they continue to grow their company. Many of these individuals are entrepreneurs and will have unique and compelling insights to share as much goes on behind the ticker. Please note, nothing in this show is investment advice and it is meant solely for educational and entertainment purposes only.
Welcome to Behind the Ticker. We are continuing our live from the Exchange ETF conference series with Weldon Rice. He is with Craft AI. They are a Korean asset manager and tech company. We talk about all the things AI is doing to help manage investments as well as some of the things we might be seeing on the frontier of AI. But more importantly, we want to talk about a couple of their ETFs. Their first, their US large cap ETF, ticker QRFT, and their large cap momentum ETF, ticker AMOM. So without further ado, please enjoy this episode with Mr. Weldon Rice.
Hey, Weldon, welcome to the show. Thanks so much. It's great to be here.
So before we get started, can you tell everybody a little bit about your background and how you ended up in your role at Craft?
Read the full transcript (50 more sections)Collapse transcript
Absolutely. So my name is Weldon Rice, head of AI ETFs at Craft Technologies. I think probably a little bit of my background, I think the uncommon would be the sort of progression of my career and life. That's probably the biggest theme. And, I started out actually not in the finance industry, worked a little bit in higher education, and I moved abroad whenever, right out of college. So I've been in Asia, living in Asia for a little over 12 years. And then, I ended up studying in finance while I was there in Korea, studying under some quant hedge fund managers and some really big players in Korea, like Goldman Sachs, head there, and really, really fell in love with finance, especially quantitative approaches to finance. So yeah, I got an opportunity to come on
Board with Craft Technologies. And I know we'll talk a little bit more about what we do in a moment. But yeah, I started out in sort of some marketing and then now heading up in the ETFs.
It's definitely an exciting time to be not only in ETFs, but also in quant AI driven ETFs. But before we get into the nitty gritty of all that, I always like to ask, what are some hobbies? What are some things you like to do outside the office?
Well, right now I have a six month old and a three month old. So I'd say my hobbies are going out to the park and, watching Miss Rachel on YouTube with my daughter and those sorts of things. But, outside of that, I've always been a big sports fan, played lots of sports in high school, and I just love watching. So it was it was great. as at time of recording, we just watched a phenomenal Super Bowl last night was incredible. So it was great to great to be in town in the US here to actually watch it.
Yeah, it was fun. It was a great game. And look, I don't envy you a minute for having to watch Miss Rachel. I went through that. I'm out of Miss Rachel. So hopefully it doesn't last too much longer for you. Yeah. So let's talk about Kraft. I know you're an AI driven manager. We'll definitely get into all that. But let's talk about can you talk about the breadth of services that Kraft offers just outside of ETFs in general?
Absolutely. So as you mentioned, ETFs is just one part of what we do. And we were actually founded in 2016. And we're founded as an investment technology company. So outside of just the ETFs, we also do model portfolios. We've run different sort of vehicles. So we're really a B2B solution provider. So we work with we're based out of Seoul, Korea. And we work with a lot of largest financial institutions in Korea starting out. And a lot of that was model portfolios or portfolio construction, helping them to launch AI powered products, their own products. And then we also do a bit of we've helped them build some robo advisors and other such things like that. So we have a bit in some of the tech side to have a little bit of footprint, but our bread and butter
Is really constructing portfolios, whether that be, security selection, mainly in US equities, or it could be multi asset, mainly using ETFs for exposure and adjusting, from a top down perspective, the asset allocation as well. So yeah, we've really expanded our breadth of what we do over the years. And in two years ago, we received an investment from SoftBank, about $140 million. So that was a that was a big change for our firm and really changed a lot for
Where we're going. So I actually had a question about that. What is an investment like that do to help kind of jumpstart the company as a whole? it probably gives you the ability to do a lot more marketing. I see you guys everywhere. You guys are doing a fantastic job of being anywhere. Can you talk or being everywhere? Can you talk about how that's really helped take you guys from maybe a smaller company to, a very institutionalized name
Today? Absolutely. So when I joined the firm, pre investment, we had about 40 people in the firm. And, we're really, we have a lot of engineers in our company, actually, majority of our employees are either researchers or developers. And, we built our AI models all in house, all vertically integrated. So we source the data, we clean the data, everything is done in house in our firm. So, what that investment did for us was the ability to one, expand our capabilities in terms of expanding into different asset classes, been able to expand even into different jurisdictions. For example, we mainly started out in US markets, but we've expanded to other Asian markets as well.
And on top of that, we, we were able to also change a bit of the frequency. So in terms of frequency of our strategies, we're looking at anywhere from intraday strategies to all the way up to about one month. So we don't really get past doing predictions more than about one month out. But yeah, we were able to expand some of our capabilities and that ended up doing some a little bit more, institutionalized strategies such as long, short and other things
Like that, that we've been able to launch in recent days. Yeah, it's exciting. And congratulations to you guys. And so let's talk about AI and kind of quantitative processes as a whole. I know AI is thrown around and there's a lot of different things that go into what actually AI is. And at a high level, you don't, you don't really have to get into specifics here, but what, what types of AI or how are you utilizing AI? Are you looking at pattern recognition, sentiment, news, all of that, none of it? at a high level, what is the firm employing in terms of trying to make these types of
Predictions? Absolutely, absolutely. So when it comes to artificial intelligence, we've just seen a huge explosion, especially with ChatGPT. And that brings with it a lot of really great, just sort of adoption and people's understanding of what it is, especially from when I first joined the company. But at the same time, it also can bring a big confusion on sort of what people are doing with artificial intelligence and especially in financial markets. And how are we using that? Even in the ETF space, we usually start out with, telling people that we're not an AI thematic fund. We're a fund that uses AI to pick the stocks in the fund. So that's usually the first point. But then, you can really drill down AI into about two different categories.
There's sort of analytical AI, which is looking more at pattern recognition, looking at what is termed in the AI industry, structured data, basically numerical data. So we're looking at, price, we're looking at different market data, different ratios. And then there's the generative AI or, AI that's looking at things like text or images and other things like that. So from our perspective, Kraft originally started more from the analytical side. We're very much from a quantitative background. Sometimes we call ourselves like a quant 2.0 in a certain sense, because what we do is we take a lot of those academically researched factors and sort of things from a lot from the quant space that have been around for years. And what we're trying to do
Is use AI to bring a more robust exposure to those factors. We're trying to look at more like risk adjusted returns, especially when we're looking at security selection in that space. In terms of what we do with the large language models, we don't employ as much on sort of textual and sentiment data. But we do have some capabilities there. And especially, recently, we just launched a new ETF with LGIA research. And one of the reasons why we partnered up with them was because they have really great capabilities in the large language models. And so that fund actually has the ability to look at sentiment data from news and incorporate that into the prediction that comes out for the fund. So yeah, we're using trying to use both technologies. And we really try to stay at the forefront and try to
Innovate and push the boundaries of what we can do with AI in the space, while at the same time, making sure it makes sense. Yeah, we don't just use AI so we can put it in the marketing materials. I'm more on the marketing side. And I totally, get that it's great to put AI and everything. But we don't we just try not to put AI in there just just because
Yeah, and you actually led kind of right into my next question, which is, how often is your process kind of building on itself? Is it using machine learning at all? And basically, the question is, you're not just letting it run wild, do you have some guardrails on it, but I'm sure you're implementing some some evolution over time.
Absolutely. So you, you mentioned a few things in there, in terms of, development of the of the AI. So, we're constantly updating and researching, we have a lot of researchers that are looking at different ways we could, employ, different metrics or different different data into what we do. And so sorry, can you go back? I got hung up on that second question.
What was the original question? The original question basically is, are you allowing the process to to are you implementing machine learning? And is that changing your process over time?
Right, right, right. Yeah. So that's actually the point I wanted to make. I remember you mentioned machine learning. And that's also a bit of an interesting topic in itself. AI, machine learning, deep learning. All these things were, before I sort of got into this sort of AI industry, were a bit foreign to me. And I think for most people, you might get confused on what the difference are between those, they're really AI is really a broad term for both of those machine learning and deep learning. We actually implore both machine learning and deep learning in what we do. And the way you can think about it is this, in terms of, traditional financial models were built on, this sort of linear models of sort of regression and sort of regressing back to the
Mean. And what we're trying to do in machine learning AI is to build a nonlinear model, because we know that markets aren't linear. All right, so we know that. So that's what we're really trying to do. And when it comes to machine learning and deep learning, just think of these as sort of like a scale. And as you slide along the scale, you start with machine learning, which are a bit more, simpler models. simpler in terms of they're less complex, a little bit easier, you can build, see a little more insights, but they have less linear, non-linearity to them. Whereas deep learning, which is what we also use in our models, that is a bit more nonlinear, also a bit harder to, dig down deeper into what actually,
Was the driving force for the decision. And so yeah, our models are constantly learning on updated data as it comes in and learning as the markets change, which is quite interesting to see.
So before we get into your ETFs, I want to talk about two of them. Does the, I just had done an interview last week with an attorney who was setting up ETFs in Europe. Is there an ETF market in Asia?
Oh, absolutely. Absolutely. There is an ETF market in Asia. There are many ETFs in Korea. In fact, Miri Asset, one of the largest asset managers in Korea actually owns GlobalX ETFs and several other ETFs around the globe. So, and they have their own locally owned ETFs that they have. So Asia has a pretty decent ETF market, I would say. Yeah. And my, actually my boss and sort of mentor in this industry, the ETF, Francis Oh, who has spearheaded a lot of what we've done in our ETF industry, has spent, his whole career, almost 12 years working in the ETF industry in Asia for Vanguard, Direction, and now Kraft and AI, which is, I always find such an interesting career trajectory to go from Vanguard to AI, especially making a stop with Direction on the way. So yeah,
Good market there in Asia as well.
So let's talk about a couple of your products. You've got, I'm just going to say Kraft, the QRFT, the US large cap ETF. From what I gathered, obviously it's AI driven, but you also include in there with human intuition. Can you explain the strategy, as a whole and
How you combine those two things? Absolutely. So what we mean by human intuition is that, and you mentioned this a bit earlier in terms of looking at guardrails and portfolio construction, there is human intuition in terms of how we look at portfolio construction for the AI. So, for example, the universe that we start with, that's chosen by humans, right? And, and sort of the factors and things that go into the model are chosen by humans. The data that goes into the model is chosen by humans, but that's about where the human process stops. We let the AI then learn the data and we, fully trust the prediction that comes out from the model. So we don't alter that. We don't go, oh, we're in this stock this month.
Really? I don't think so. We better, we better, dial that down a bit. We don't do that. We let, we trust the process, but at the same time, you mentioned this a bit earlier, we do have guardrails in terms of, in terms of how much concentration we can get into different sectors. And so, yeah, we, we definitely do. And we have some, risk controls within the models to, to help, help that out as well. So yeah, we, we really try to follow some, the investment process of, human managers, but we're trying to let AI have a different look at, at what the data is telling us and really bring a new approach, I think, to,
To portfolio. So you're investing in, in large cap U S securities or large cap U S equities only about how many names can be in the portfolio at any given time.
So for QRFT, that can be about 300 to 320 around there. So that fund is a really a well-diversified fund. And we, we, we made that really to be an enhancement on sort of a, a traditional large cap passive exposure. we're really trying to, I, we kind of call it like say that we're long-term investors, but at the same time we have a short-term, view. So we really try to make adjustments in the short term on a monthly basis, uh, trying to, find sort of patterns in the market to find little inefficiencies here and there. And then over the long-term build up a really robust risk adjusted return. And hopefully that, we believe
That's going to be, bring alpha over time. So with 300, 320 names, how does the portfolio weighting work? Is that a machine driven process to kind of pick out, and rank maybe the best ideas or can you just talk about portfolio weighting? Absolutely. So the first part of the
Process, uh, is a bit more looking at our factors. So we use traditional academic factors as well as some of our own proprietary factors. And then we, score the securities based on that. And then we also have a model that does the weighting process as well. So we, we, we look at, okay, here are the, here are the scores from the factors. And we look at, put that in another model with other market data, as well as some of the fundamental data from the securities. And that will give, uh, sort of its prediction for what's the best stocks that we feel over the next four
Weeks have the best price appreciation. So how, um, how is a portfolio rebalanced or traded? Is that
A monthly rebalance? Did I, did I hear that? Yes, that's right. Yeah. So we do a monthly rebalance
And we, yeah, for, for that fund. So I know being active, we run an active strategy. You guys are running an active strategy. Can this, can, uh, QRFT get defensive based on your signals or does it always kind of stay fully invested or does the, the defensiveness of you're seeing weakness in the market kind of direct in which assets it may be going to? Yeah. because it's equities,
There's a, only a certain amount of defensiveness you can, you can get, but well, because we're looking at different factors in the market, for example, when the market does get a bit more defensive, we see that, value factors and quality factors tend to be the, the factors driving the model a little bit more. Whereas, in, in, when the market's really growing, we start to see a bit more on momentum, uh, and some of these other growth factors, uh, that come in. So it can get a bit, a bit more defensive and it really adjusts itself over, over time.
So speaking of momentum, you also have AMOM, uh, which is your large cap momentum ETF. So what is the biggest difference between the U S equity large cap and this one, which is the momentum version?
Right. So in this case, we're, we're sort of singling out one factor, uh, the momentum factor, whereas QRFT, it's looking at multiple factors across, there's many, many factors there. So that's one key difference. Uh, the other is the number of holdings. So, uh, QRFT is very well diversified. Uh, I call it kind of a slow mover because, it's got a lot of, a lot of holdings in there. AMOM is, uh, is a much more, uh, we, we like to say, uh, cause we're from Korea. We like spicy food. AMOM is a lot more spicy. Uh, we only have 50 holdings there. Uh, we can get some pretty high concentrations in, in sectors, uh, and in stocks. we have some guardrails there,
But we can go up to 50% in a sector and we can get somewhere around nine to 10% in a holding at some point. Uh, and I think recent market conditions have sort of, been really interesting to see, uh, because, the market is very concentrated right now. And so we're actually seeing our momentum fund, really get some pretty, uh, pretty big concentration
In some of these holdings. It's interesting. So is it also just to compare the two again,
Is it also traded on a monthly basis? Yes. Yes. It is also a, just like your RFT on a monthly basis.
Interesting. So I know you guys have a lot of other products and I'm going to have you or somebody else from your team back on, cause I know you're going to keep launching ETFs, but let's talk broadly. Um, as we kind of close out, how should advisors, if you're a registered investment advisor, how should they be thinking about implementing AI driven strategies inside their portfolio? Like, do you see them as a holistic solution or more of like a satellite exposure? They should be tagging or adding along maybe some more traditional passive holdings. Yeah, absolutely. I think,
Our, our view is that, AI is, we definitely believe in the technology. We believe in what we're doing, but at the same time, we understand that, uh, AI is not perfect. And so, we believe that AI is a great compliment. when you think about what artificial intelligence is doing with the data, it's really taking a very different view of the market and data than a human manager would. So we believe this provides a very interesting diversification effect for, advisors to look at. And this really compliments a lot of either, QRFT, I think could be a good compliment to maybe a more of a core exposure, passive exposure. If you're looking for a little bit of active enhancement, uh, to a more of a passive
Exposure to sort of a large cap, I think it's a great compliment to, to be sort of a sleeve there to give an active, uh, data-driven systematic, but flexible approach. Uh, and then AMOM is, is a bit more of a satellite stylistic fund. That's going to give you, uh, an ability to be a bit more of like trend following of the market. So instead of just taking that long-term view that, okay, the market's going to go up over time, let's take some, bets on the market of what, what we're looking at currently and trend, follow some of the trends that are going on, uh,
In the market. So kind of a last question and it's, it's, it's AI driven again. It, I agree with you and I have the same sentiment as you that AI is interesting. It can do a lot of things, but it can't do everything. Um, you have a lot of engineers on your team. At what point do you think in total prediction where nobody's going to hold you to this, you think that maybe the human side is going to take a step and maybe sit in the back seat a little bit and, and just let AI make all of the decisions?
That's, uh, we believe that that future is very possible, maybe not for every, uh, but we definitely see that this is really, I think the future of, of where a lot of portfolio construction is going to happen, uh, going forward. Uh, but at the same time, uh, there are so many different reasons why people invest in different themes that they might want to, want to cover. So AI is not going to cover every single thing, but we believe that, uh, majority of, portfolio construction and, and things going forward is actually going to be, have some part in AI, whether it be, the AI actually making the stock selection or whether it be, analysts or teams using AI for their research, uh, and coming up with, uh, some ideas on,
On, on, on their universe or their investment thesis. Uh, we definitely see AI as a huge part
Of that going forward. Well, Weldon, I can't thank you enough for joining us. I really appreciate it. Before I let you go, where can people learn more about you and learn more about craft technologies?
Absolutely. Well, you can always find me on LinkedIn, uh, and craft technologies. You can check out, uh, craft AI ETFs.com or craft tech.com. And by the way, craft is with a Q, not a C, uh, and actually craft stands for quantitative craftsmanship. If you're, if you're curious about that. Uh, so craft tech.com or craft AI ETFs.com, you can find us. We're always happy
To, to reach out. Well, again, thanks for being here. I hope you enjoy the rest of your week. Appreciate it. Thank you very much.
Thank you.
Daily Market Intelligence
The Signal
Brad Roth's daily market brief — systematic signals, ETF positioning, and what the data is actually showing.
Subscribe Free →