Frontrun
The goal
Frontrun is a trading platform built for people who live in the markets. The job had three parts: define the product strategy, set the design strategy and build a clickable prototype fast. On top of that, lay the foundational AI strategy for how AI should live inside Frontrun and become a real differentiator.
The challenge
Traders drown in information. Feeds, charts, wallets, group chats, scams. The risk with adding AI to that is making it a gimmick. So the rule was simple. AI has to be a tool that makes a trader more efficient: find the right information faster, comb through the trends, cut the noise. And the whole prototype had to come together in a matter of weeks, not months.
My role
I owned the product requirements, the user strategy and the voice of the customer, and I defined how AI shows up inside the product. The bar for the AI was simple: help a trader read the context and the market trends fast, then act. I set the product and design strategy, owned the AI development and drove it to a clickable prototype.
How I approached it
Decide what AI is for, then design around it. In Frontrun the assistant earns its place by doing a trader's grunt work. Summarize the unread messages. Flag the scam messages. Pull up a token's chart the moment you ask. Surface the trend instead of making someone dig for it.
To hit the timeline, we went from strategy to a clickable prototype in weeks, a proof of concept the client could put in front of investors and react to.
AI earns its place when it makes a trader faster, not when it sounds clever.
The result
A clickable prototype with the AI assistant as the differentiator, ready for investor conversations, plus a foundational AI strategy for where the platform takes it next.
