Infinite Quant — The Signals Program

The vision

Most quant operations treat alpha discovery as an internal function — a small team, proprietary models, closed-loop research. That works at certain scales, but it leaves an enormous amount of human-generated edge on the table. Independent traders, scattered across exchanges and prop firms and timezones, are constantly noticing things no internal team has time to find.

IQ’s bet is that those traders will contribute their best signals to a shared platform if the incentives are honest, the tools are genuinely useful, and the upside is real. Signals from contributors are validated, ranked, and — when they perform — folded into the platform’s broader research stack. The AI side of the system learns from the human side. Over time, the line between “AI-generated” and “human-generated” alpha blurs into a single, compounding research output.

For investors, the structural value is straightforward. In most quant funds, research is a fixed cost — researchers draw salary whether or not their work produces. IQ pays contributors only when their signals resolve as winning trades. Research spend scales with output, not with headcount. By absorbing validated signals into the AI stack, IQ also builds a data moat that compounds with every cycle. The Signals Program is, functionally, a research department that scales without headcount and pays only for output.

What participants are joining

Joining the Signals Program is not an investment, not employment, and not a funded-trader arrangement. Participants don’t contribute capital, don’t trade IQ’s accounts directly, and don’t receive equity, salary, or any economic interest in Delta Prop Shop LLC. They contribute signals, and they’re paid when those signals resolve as winning trades on platform-managed accounts.

There’s no application fee, no skill prerequisite, and no prop firm requirement. Anyone can join. Earnings begin after a warmup period of ten resolved calls — wins or losses; washes don’t count toward the count. At one submission per day this typically runs two to three weeks, depending on how many of your calls resolve cleanly within their horizon. During warmup you submit on the same cadence, get the same agent walkthrough, and see each call graded in real time. The moment your tenth resolved call is graded, your hit rate is computed and you’re classified — Tier C (30–40%), Tier B (40–50%), or Tier A (50%+) — and your winning trades start counting toward the monthly pool. A hit rate below 30% keeps you in warmup; random direction on this target hits roughly 35%, so clearing the 30% floor is the bar for showing edge.

Each participant receives access to IQ’s research and coding agents, the backtest infrastructure, validated AI-generated strategies, and a browser extension that brings IQ’s tools onto third-party trading platforms. Whatever a participant builds with these tools — strategies, backtests, code, behavioral patterns — they keep the practical ability to use elsewhere. Trades placed on external platforms can be submitted back to IQ as signals; if those signals resolve as winning trades on platform-managed accounts, they count toward the participant’s earnings.

The deal is straightforward. As long as you’re submitting winning trades, you share in the program’s monthly payouts and your use of IQ’s AI tools is free — for trades on IQ or anywhere else you trade. If you stop contributing, you keep access to the tools but pay for your own usage.

How the compensation works

Net monthly trading profits split two ways: investors and the manager. The investor/manager split typically lands between 50/50 and 65/35 in favor of investors, and shifts month to month as new investors join under different profit-sharing terms that depend on the fund’s performance.

The manager share is then split between the AI system and the trader pool based on measured alpha contribution. There’s no AI floor: if traders generate all the alpha in a given month, the entire manager share flows to the trader pool. Within that pool, payouts are pro-rata by winning trades, distributed on the last business day of each month.

A winning trade is a call where price moves in your submitted direction to the qualifying target before either the stop-out level is hit or the call’s time horizon expires — whichever comes first. It’s a fixed asymmetric barrier, graded automatically against the mark price on the platform-managed mirror account. If the stop hits first, it’s a loss (you owe nothing). If the horizon expires before either side, it’s a wash and doesn’t count either way. Fees, funding, and slippage on the mirror are absorbed by the platform — your grading is purely a directional question, “did the call work?”

To make this concrete: in a month with $100,000 of net trading profit and a 60/40 investor/manager split, the manager has $40,000 to allocate. If 60% of that flows to the trader pool — illustratively — the trader pool is $24,000. A participant with 12 winning trades in a pool of 40 total receives 12/40 of $24,000, or $7,200.

Knowledge flow and the data moat

What you submit becomes part of IQ’s research stack and may be replicated system-wide. Continued earnings depend on continued contribution.

For investors, this is the compounding mechanism. Each month, the AI side of the system absorbs more validated signal patterns, becomes more capable of generating its own variants, and reduces the fund’s dependence on any individual contributor. The Signals Program is both a source of present alpha and a training pipeline for future alpha.