Custom-trained AI · radical transparency

Autonomous trading strategies,
built on a custom-trained AI.

A state-of-the-art time-series foundation model, trained on 1.5 billion US-equity bars, drives a disciplined rules engine that trades unattended. We sell better-shaped equity exposure — not magic, not alpha — and we show the receipts.

Three profiles, one engine

Each profile is the same AI forecaster at a different risk setting — vol-managed, balanced, or concentrated. Pick the shape that fits you.

Steady

Lower risk · vol-managed

US mid-cap exposure, volatility-managed for shallower drawdowns. For copiers who want equity upside without the gut-punches.

Core

Balanced · index-relative

SPY-like exposure with regime-adaptive name selection. The mainstream profile, built to be compared against an index fund.

Edge

Higher risk · concentrated

Concentrated, high-conviction positioning aimed at long-run compounding. Deeper drawdowns are the price of admission.

How it works

A custom-trained time-series foundation model forecasts short-horizon moves across a mid-cap universe. A deterministic policy turns those forecasts into target positions, and an executor routes them to a copy-trading broker — 24/5, no human in the loop. The intelligence is a trained forecaster plus a rules engine, not a black box making per-trade bets.

Why trust it

Transparency is the product as much as the returns are.

Backtest and live, side by side

When the track record is live, you see the in-sample backtest next to real results — including the gap.

Pre-registered kill criteria

Each strategy carries published rules that auto-halt it if live performance breaks down. No moving goalposts.

Honest framing

Better-shaped equity beta, clearly labelled. We don’t claim to beat the market by picking winners.

Pre-launch. BitterQuant is in paper-trade validation. A live, independently-tracked performance dashboard is coming here. No performance figures are published yet — any future figures will be clearly labelled as backtested or hypothetical, with past-performance and risk-of-loss disclaimers. Nothing here is investment advice.