Leo Martinez Was Told to Quit Math. Then He Started Trading $30 Million a Month.

He dropped out of college, lost $20,000 in two days, taught himself to code before AI could help, and ended up co-founding an ad-tech company valued at $75 million. Here’s how.

Leo Martinez’s high school math teacher signed his yearbook with seven words: I hope you never take another math class again.

“It was hurtful,” Leo says, “but it didn’t tell me anything I didn’t already know. I was already pretty defeated when it came to that. But yeah—I thought it was kind of insane.”

A few years later, Leo would be writing trading algorithms that moved $30 million a month through crypto and equity markets—without a degree, without knowing anyone who could code, and without any of the credentials that are supposed to come first. Today he’s co-founder and CTO of Gravity, an AI-native advertising network valued at $75 million. The thread connecting every chapter of his career isn’t talent or luck. It’s a method: try something, measure the result, adjust fast, and go again.

The short version of Leo’s first career: he dropped out of college at eighteen, got rejected from Equinox, gave away over 100 free personal training sessions at the Four Seasons Hotel Vancouver—paying the facility fee out of his own pocket—to build enough experience to reapply. A month later Equinox hired him. Within three months he was a top-selling trainer. Over four years he reached the highest commission bracket and was named Trainer of the Year at Equinox Canada.

He left at 22, tried trading cryptocurrency on instinct, and lost roughly $20,000 in two days. He studied how professional firms operated, saw algorithms instead of intuition, and taught himself to code from scratch in 2020—before AI tools existed to help. It took weeks before he could do anything basic. He kept going. He built proprietary market-making algorithms that generated over $30 million a month in profitable trading volume, and his systems later became part of quantitative infrastructure alongside an investment group deploying roughly $200 million across crypto and equity markets.

Every transition followed the same cycle. Identify the gap. Close it as fast as possible. Use real results as the proof. That pattern is what makes the next chapter make sense.

In 2023, Leo reconnected with Zach at a crypto conference in Mexico City. Zach had spent years in ad-tech and had built Flax Labs, a marketing agency focused on measuring the actual profitability of an ad unit—not just revenue. The standard metric in advertising, ROAS, tracks top-line return on ad spend. It can make a losing campaign look like a winner. Zach had built software to measure profit at the ad-unit level and adjust budgets in real time.

The conversation wasn’t a pitch. It was two builders comparing how they thought about systems. They kept talking—months later in Dubai, still riffing on the same thread. In September 2024, Leo joined Flax Labs as CTO. Working there, he started seeing how closely advertising infrastructure resembled the trading systems he’d already built: two-sided marketplaces, real-time decisions, incomplete information, and a premium on speed.

In January 2025, both stepped away from Flax to start what would become Gravity. What followed was eight months of intensive iteration. Testing ideas, getting feedback, deciding fast, and moving on.

The first thing they had to figure out was how to work together. Zach brought deep ad-industry experience and ambitious timelines. Leo brought a grounded technical perspective. The dynamic worked because both sides served a purpose.

“Zach’s ambitious timelines and goals accelerated our growth,” Leo says. “As we were iterating, I was able to keep us grounded. His big goals combined with my groundedness helped us stay on track and move quickly.”

The original idea was to automate marketing as a whole. The thesis was appealing: if you could democratize access to high-quality marketing, the best products would win rather than whoever had the best growth hacker. They built toward it, tested it, and realized the business model was difficult to sell. So they moved on.

That became the rhythm. Try an approach, get real feedback from the market, decide whether to continue or pivot, and keep the cycle short. Leo estimates they changed the plan at least ten to twenty times over those eight months. Each iteration wasn’t a failure—it was information.

The ad-network model emerged from that process. Once it clicked, the pace shifted from searching to building.

Leo built the entire initial version of Gravity’s ad network from the ground up, before they hired anyone. The architecture reflects his background in quantitative trading: high-performance, low-latency, and built to make real-time decisions at scale.

The core of the system is a relevancy-based matching engine. On every user query to an AI platform, Gravity identifies the most relevant advertiser. Leo built the machine learning models that score relevancy across queries and advertisers, and those scores also inform pricing for each ad unit. Relevancy was a priority from the start—a relevant ad is less intrusive inside an AI conversation, and at its best, it’s genuinely useful. If a user has a problem and an advertiser offers a relevant solution, the ad becomes part of the answer rather than a disruption.

The second major challenge was latency. Gravity generates a unique ad for every impression using a large language model—no two ads are the same. That means the system has to run inference, match an advertiser, generate custom ad copy, and deliver it fast enough that the user experience isn’t disrupted. Leo applied techniques from high-frequency trading: hardware optimization, code-level performance tuning, and infrastructure designed for speed.

The premise behind Gravity is straightforward. AI tools shouldn’t require subscription fees to be accessible, and advertising can fund that access—the same way it funded the internet. Google, YouTube, and Instagram are free because of ads. Gravity’s bet is that AI follows the same trajectory.

“We put ads inside of AI,” Leo says. “The reason is to subsidize the cost so users can use AI platforms for free, just like how Google search is free because of advertising. Otherwise, we’d all have to pay a monthly subscription fee, and probably no one would do that.”

Gravity raised its seed round in October 2025 at a valuation of $75 million. The company now serves millions of ads per day to AI platforms. Leo works alongside the team in Silicon Valley, drawn by the same logic he’s applied throughout his career: go where the feedback is fastest and the competition is highest.

For more information, visit trygravity.ai