Listen to our latest episode featuring Stratwise and learn why a strategy built to survive first and earn second can turn market volatility into income without leverage.
In this episode, Nertila from Trading Strategy talks with Iaroslav from Stratwise. Stratwise is an AI-driven crypto asset manager that runs non-custodial, spot-only strategies, including public vaults on Hyperliquid, using a machine-learning engine to trade volatility. Its site highlights a dynamic adaptive grid that forecasts volatility and squeezes and re-optimizes in real time.
Play video: Episode #16: Stratwise- View Stratwise ratings on Trading Strategy
- Watch the Stratwise episode on Youtube
- Listen to the Stratwise episode on Spotify
Any questions? Join our community Discord server.
Transcript:
Welcome to TradingStrategy.ai, the podcast where you can learn the essentials of DeFi trading and new worlds. This is your go-to show for automated trading strategies and market insights. Welcome to the conversation. I'm your host, Nertila.
Nertila: Hi, everybody. Welcome to today's TradingStrategy.ai podcast. Today we have another guest. His name is Iaroslav, and he's a co-founder of Stratwise.ai. Hello Iaroslav, and welcome to the show.
Iaroslav: Thanks for having me. A quick disclaimer: everything I say is from live trading and is not investment advice.
Nertila: Great, let's get started then. You describe yourself as an operator rather than a trader. What did you actually do in crypto before Stratwise?
Iaroslav: I got into crypto in 2017 as a student, and finished my economics degree at Plekhanov University in Moscow after that. So crypto came first, the diploma second. I've worked on a trading desk, ran the crypto department of a large CPA network company, started a crypto marketing agency, and I'm currently a co-owner of a neobank in Asia.
All of that is the business side, not the trading side. What it gave me is a network of people with money, and a very close view of how they lose it. That's the short version.
Nertila: What was the specific problem you kept hearing from people with capital that made you start an asset manager?
Iaroslav: Things blow up. Most trading products a normal investor can buy have one of two problems: they use leverage, or they keep adding to a losing position with no limit. Both look great on day one, but one move wipes the account. The serious quant funds that manage risk properly, on the other hand, won't talk to you unless you have eight figures.
So there's a gap: people with $200,000 to $5 million who want something run like a fund but as easy to plug into as a bot. That's what we built.
Nertila: You run the business side and your co-founder runs the AI. How does a CEO who didn't write the models sell them credibly?
Iaroslav: I don't sell the models. I sell the proof. My co-founder, Vladimir, runs all the AI and ML work at our company. The team comes from MSU and MIPT, and they published two peer-reviewed papers at ACM 2025 on how the strategy works. We've only been live for about a year, so to be honest, peer review is a substitute for a long track record.
My job is to make sure anyone can check it and say, "All good." Our full due diligence questionnaire is public, our vault is public on-chain, and investors get daily numbers. I don't need to have written the models myself. I need to make them checkable. That's the selling point.
So that's my setup: my network, my team's research, and a strategy built to survive first and earn second.
Nertila: Very interesting. Explain the strategy to someone who has never heard of a grid. What does the machine learning actually decide?
Iaroslav: Instead of buying all at once, we place a ladder of buy orders below the price. When the price dips, we buy a bit; when it bounces, we sell for a small profit, around 2% per round trip. We do this many times a month, turning over about seven and a half times the account size, and we trade spot only, no leverage.
If we do trade perps, it's at 1x leverage, the same as spot. The AI doesn't try to guess where the price goes; that's where most crypto AI fails. It looks at how volatile the market is, and at liquidation squeezes and other sharp moves. From that it decides how tight the ladder is, how big each order is, and where to place them.
Volatility isn't our enemy; it's what we earn from. That's how the model works.
Nertila: Why non-custodial? What did you give up by never holding client assets?
Iaroslav: Honestly, non-custodial isn't a selling point anymore. It's the basic standard for a lot of teams on the market. You open a sub-account on your own exchange, give us trade-only API keys, and you can stop any day, with no lockups. What we give up is convenience: we can't pool money, so there's more admin work and capacity is limited per account.
What the investor gets is that the only question left is what's actually running behind the API key. Most products are a black box with a nice return number. Ours comes with the research, the questionnaire, and the real drawdown history, not cherry-picked data. Fees are 40%, which is normal for the market, and of course there's a high-water mark.
Nertila: Walk me through a normal week. What does the CEO of a systematic fund actually do when the system trades itself?
Iaroslav: First, investors: onboarding, reporting, and answering the same 50 due diligence questions in 50 different orders. Then the fund structure: lawyers, administrators, the really boring things. And I sit with Vladimir's team going through model updates and risk alerts.
We have a lot of calls during the week about how things are working and how we can improve the technology. And of course content, because that's how we grew. It's much less exciting than it sounds. Most of my week is reading logs and PDFs, though sometimes we do posts and PR work.
Nertila: I want to move to risk management. You run spot only with no leverage. Isn't that leaving money on the table?
Iaroslav: Any grid bot will show you a better backtest than ours if you run it at 3x. We chose not to. Volatility is our income; leverage turns that income into liquidation risk. Our risk management has three layers. First, no leverage, so we can never be liquidated. Full stop. Second, automatic rules that shrink position size when we're in a drawdown and cap how much capital any position can use.
And third, the reverse grid, which I think is your next question.
Nertila: What is the reverse grid mechanism, and why is activating it still a human decision?
Iaroslav: When all our buy orders are filled and the price keeps moving against us, the reverse grid opens a position in the other direction to stop the bleeding and help the account recover. We have a 20% stop on every account, and the reverse grid kicks in at around 12%. Every investor agrees to these numbers before they start.
Right now, a human still has to approve turning it on, and that happens within a day of the alert. I'd rather tell you that than have you read it in our documents. It's a big decision with a real cost for the investor, so we wanted a person in the loop while we collected live data on it. We're automating it now.
Everything else already runs on its own, fully automated. But for this part of the strategy, we really need the investor's decision.
Nertila: How did your strategies get through the October 10, 2025 liquidation event?
Iaroslav: We had zero liquidations, because there was nothing to liquidate. Our system buys into sharp drops, so it effectively captures the recovery instead of getting killed by the crash. We earned around 3% in a single day on October 10, 2025.
The worst day of the year for the industry was a green day for us. And that's the whole product in a nutshell: it helps users manage risk much better. This isn't plain futures trading or a standard grid. Events like that will never hurt us. But trends can.
Nertila: Where do you think capital risk will accumulate over the next cycle?
Iaroslav: Capital now goes to managers who can prove they made money live, through good and bad markets. Nice backtests or one good year don't sell anymore. Risk is building up in two places. First, leverage built on top of stablecoins and restaking, where nobody really knows what the collateral is worth in cash. I think that's a huge risk.
Second, exchanges that are also the custodian and the lender. Everyone stress-tests their strategy; almost nobody stress-tests the place their money sits. We're also really worried about AI risk for the whole crypto industry: AI technology is growing day by day, and so are the exploit risks for DEXs and CEXs. That's what we're thinking about right now.
Nertila: I want to stay on that story, the worst day in crypto being a good day for you, but in reverse. Your worst drawdown was a rally, not a crash. Can you tell us more?
Iaroslav: In May 2026, one account was set up short HYPE. One of our investors put $1 million in HYPE, and it's important to underline that it was a HYPE-denominated strategy. Then the ETF launched, and HYPE went up for two weeks straight. No pullbacks, no corrections. We were down about 20% at the low point.
It was the biggest drawdown we've ever had, and honestly, it was really stressful. We turned on the reverse grid with the investor, and the account came all the way back and is in profit now. The lesson was simple: our strategy needs the price to move back and forth. A crash is fine; we buy into it.
A straight line in one direction, with nothing to buy or sell against, is the one thing that hurts. So we tell every investor that before they start, not after it happens. Most of our investors use our strategy as a way to diversify their portfolio.
Nertila: Your strategy averages down. How is that not martingale?
Iaroslav: Simple. Martingale keeps doubling until the money runs out. Ours has a hard cap on how much capital any position can use, and order sizes shrink or grow based on how volatile the market is. It can't run away. It's also important to note that we rebalance capital between assets in our multi-asset strategies.
The other difference is who sets the steps. In most bots, you type in a number once and it never changes. In ours, a model looks at current volatility and sets the step size and the cap for you, and it learns from live data. If you're ever looking at a grid product, ask one question: where's the cap?
If they can't answer it in one sentence, walk away. It's that easy.
Nertila: Why did you refuse to give investors read-only API access, and what did you offer instead?
Iaroslav: We do offer it now. Before, with read-only API access, anyone could see our live orders, copy the strategy and trade alongside us. That was a huge risk for us. But now that we're in the final stages of incorporating a private fund, we're ready to share it with our investors.
We're also providing not only API keys but our wallets, for example on Hyperliquid, for real data transparency, because copy trading is no longer our problem.
Nertila: If you could change three things about the digital asset management ecosystem, what would they be?
Iaroslav: One: track records that live on-chain, so nobody can fake them. Two: exchanges shouldn't also be your custodian and your lender. Three: make it cheaper for small managers to get regulated. Right now, getting regulated costs so much for a manager under $50 million that they stay in unregulated vaults or don't exist at all.
There are a lot of good teams that allocators can't touch. That's part of why we're building a proper fund ourselves. It's expensive, but now we can handle it.
Nertila: What's next for Stratwise over the next twelve months?
Iaroslav: Three things. The fund is almost incorporated, so investors who can't do managed accounts get a proper vehicle. We're fully automating the reverse grid. And we're adding more exchanges and more assets.
We're also researching new types of strategies. Our team is working on a collateral strategy that lets investors put up Bitcoin as collateral. We have really good backtests on it, and I think we'll start live trading tests next week. We're also looking at market making, part of a family of delta-neutral strategies.
We're already testing it on our own balance and our own exchange accounts, but it's still in testing. Two months of track record isn't a lot for investors.
Nertila: Last question. If a listener remembers one thing from this conversation, what should it be?
Iaroslav: Ask every manager two questions: How much leverage do you use? And where is the cap when you average down? If the answer takes more than one sentence, the product is built to look good, not to survive.
Nertila: Do you want to mention where people can find you? I see here that the full Stratwise due diligence questionnaire is public.
Iaroslav: Yes. You can find our information at stratwise.ai. The full due diligence questionnaire is public at stratwise.ai/ddq: 55 questions, no email wall. And follow Stratwise AI on X. Thank you so much.
Nertila: Thank you so much, Iaroslav. It was a very interesting conversation, and thank you again for joining me on the show.
Iaroslav: Thank you for inviting me. I'm really glad we had this talk. I appreciate it.