Steve Lacy, founder and CEO of Cotes AI in Troy, Michigan, joins Sunny Ray to explain how he applies signal processing to financial markets. A Stanford dropout who worked at early Silicon Valley companies like Affymax and Affymetrix, Lacy grew up around control systems and later noticed that decomposing price charts spectrally revealed patterns similar to control system responses. He built Strike Stock AI to test this on equities before pivoting to Spectral Bot, which analyzes Bitcoin fifteen minute up down markets on Kalshi using aggregated trade data across multiple time scales. The engine has now processed over twelve thousand sessions, and Lacy is focused on proving a clean, unmodified live trading track record before opening broader access. He discusses why he chose prediction markets over equities initially, why he avoids handling customer funds or giving financial advice, and his plans to eventually return to equities and possibly relocate the company. The conversation covers fundraising, product philosophy, and the surfer analogy Lacy uses to describe hidden market waves.
A Stanford dropout turned signal processing engineer explains how he decomposes markets into hidden waves to predict price moves.
What problem is Cotes AI solving for retail traders?
Retail traders get tools like MACD and RSI that lag the price chart by ten to twenty minutes, while firms like Citadel make millions of trading decisions on FPGAs in the time it takes to click a trade. I approached this from a low level signal processing standpoint to figure out how to be much faster using one minute or five minute data instead.
What put signal processing and financial markets in the same room for you?
I started looking at stock charts and recognized them from control system theory homework, where you derive equations in the time domain that take pages of calculus, but in the frequency domain it becomes simple algebra. When I decomposed something like an Ethereum breakout spectrally, it looked exactly like a control system response, similar to a cruise control or air conditioner.
When did you first believe price is a signal rather than noise?
It started with a breakout I studied after a presidential tweet, where I saw phase lock about forty five minutes before the actual breakout and it stayed locked until an hour after. Think of a surfer looking for a wave, if you could strip away the noisy waves further out in the ocean, you would see the big wave coming before it rises near shore.
Why start with prediction markets rather than the stock market you already knew?
Investors wanted three months of live results on a fifty thousand dollar bank I funded myself before considering investing, which felt like a huge gamble. Prediction markets were newer and seemed more investable, and as a father of five with a young adult child, I wanted to bring institutional level risk control to younger people betting on these zero sum binary markets.
Can you explain Spectral Bot in one minute, what goes in and what comes out?
We take BTC trade ticks from Coinbase, aggregate them into five second bars, roll those up to five different time scales up to fifteen minute bars, then feed everything into ML that sorts through about eight hundred to nine hundred spectral fields across all time scales. It builds a model and outputs an up or down signal with a confidence level, like eighty five or ninety percent.
How do you sell a real edge without promising one, given you're not a financial adviser?
We publish a signal and never handle anyone's money, though we have full auto trading internally that we don't sell. It's based on historical data, similar to saying a plane will leave the ground because of physics, but historical results don't guarantee future results.
What has to be true in the next ninety days before you open the doors past early access?
We only settled on a fixed strategy and execution setup this past weekend after constantly changing parameters and models. I want to see the bank grow for two to four weeks on that unchanged setup so we have a clean track record, because you don't want to give people access to something that could lose their money.
Building something daring? Sunny talks to founders like this every day. Fifteen minutes to see if your story belongs on the stage.
Claim your pre-interview