Breaking stories and in‑depth analysis: up‑to‑the‑minute global news on politics, business, technology, culture, and more—24/7, all in one place.
More than 230 teams entered the WEEX AI Trading Hackathon, but only 37 advanced to the finals, where their artificial intelligence trading strategies faced live market conditions. Among them was Janet Ekka, founder of Smart Money Tracker (SMT), a solo-built system integrating whale tracking, sentiment analysis, technical indicators, and a multi-persona decision architecture designed to withstand extreme market volatility.
WEEX, a global cryptocurrency exchange with over 6.2 million users across 150 countries, hosted the hackathon as part of its commitment to advancing AI-driven trading innovation. Ekka detailed how her system functions in live markets, the critical lessons learned from navigating real-time flash crashes, and why, in her view, survival must always accept precedence over profit.
At the core of Smart Money Tracker is a multi-layered architecture that monitors whale activity, analyzes order flow, assesses market sentiment, and incorporates technical indicators. Each module operates as an independent analytical “persona,” generating structured arguments rather than isolated numerical triggers. According to Ekka, this approach avoids reliance on a single factor and prioritizes conviction over frequency.
The system culminates in a decision-making engine, internally referred to as the “Jury.” Its role isn’t simply to aggregate signals, but to weigh confidence levels, validate alignment across personas, and determine whether market conditions are structurally stable before deploying capital. Ekka explained the system using an analogy to crossing a busy street: “Multiple observers gather information, but a trusted decision-maker determines if it’s safe to proceed. It’s not ‘if X > 0.7, sell.’ It’s understanding why whale distribution coincides with aggressive taker sell orders and what that context means.” She argues this reasoning layer differentiates AI from simple automation.
The recent flash crash served as a real-time stress test. While some strategies attempted to trade the volatility, Smart Money Tracker retreated. “By design, it’s a coward,” Ekka stated. The system reduces exposure when persona alignment weakens or volatility exceeds predefined thresholds. If signals conflict, execution pauses entirely. Under extreme conditions, the AI can suspend trading for hours. During the crash, Ekka implemented dozens of refinements, strengthening flash crash protection, raising confidence thresholds, and adjusting internal weighting logic.
The event too prompted a shift in signal hierarchy. On-chain whale data demonstrated 80-90% confidence during the turbulence, leading Ekka to believe these signals deserved greater weight.
Following a drawdown to approximately $4,100 in equity, Smart Money Tracker entered recovery mode. The path back toward $10,000 required approximately 7-8% daily growth—a mathematical challenge demanding precision rather than reinvention. Ekka implemented three upgrades: a profit-lock mechanism that secures gains of 1-2% instead of waiting for 15% targets; a “Fear Shield” to protect profitable positions during extreme fear and greed conditions; and a hard cap of three concurrent positions to reduce fee bleed and increase conviction per trade. “The strategy that delivered 566% in the qualifiers still works,” Ekka said. “What broke wasn’t the signal quality—it was the position management.” Version V3.1.64 represents the most refined iteration to date.
Smart Money Tracker is built on free cloud infrastructure, open-source tools, and public APIs. “You don’t need a Bloomberg terminal or a PhD in quantitative physics,” Ekka noted. “The barrier to entry for AI trading is zero. The barrier to great AI trading is sleep deprivation and stubbornness.” She urged other developers to release 80% versions of their projects, capture incremental gains, and allow compounding to do the rest.
The WEEX AI Trading Hackathon differed from many others by utilizing real capital. “If your bot opens a fictional BTC position of $31,000 with 18x leverage, you feel it,” Ekka explained. A 1% move represents an 18% equity return—or loss. Code written at 2 a.m. Was no longer theoretical; it directly determined financial outcomes. The transparent leaderboard left no room to conceal weak risk management. Ekka affirmed a fundamental truth: intelligent systems are defined not by how aggressively they trade, but by how well they endure.
