Can a signal remain traceable from research to paper execution?
The same signal should have clear data provenance, a consistent backtest interpretation and risk controls before it is considered for a paper workflow.
Projects/Helix
A research system
with an audit trail.
Helix connects signal research, backtesting, portfolio controls, validation and paper trading. The design goal is to make each step inspectable before a result reaches a decision.
- My role
- Design, implementation and validation
- Since
- June 2026
- Markets
- US, China A-shares, Hong Kong
- Code
- Private repository
Validation gate · this snapshot
Signal parity · passRisk · passDeflated Sharpe 0.21 · fail
The system rejects its own book.Real screenshots of the Helix web front. The terminal shows a saved snapshot (27 Jul 2026, 50 liquid US stocks): an in-sample backtest whose combined book fails Helix’s own deflated-Sharpe gate. Its numbers are not performance claims.
System design and validation.
I designed and built Helix: a shared signal core, a factor lab, several backtest engines, a portfolio allocator with risk limits, validation gates and a simulated paper-trading loop. It builds on others' work: Qlib for one ML sleeve, and a factor engine and factor library vendored from HKUDS Vibe-Trading under the MIT licence.
Checks along the chain.
The local system documents signal parity checks, embargoed walk-forward research, cost-aware backtests and run records keyed by code, configuration and data. The diagram above is a map of those components, not a performance chart.
Paper evidence has a boundary.
Point-in-time universe masking is deliberately not routed through strategy backtests, so absolute returns are overstated by about five points a year while relative comparisons hold. Delisted price coverage is incomplete, and the recorded paper history includes replayed days. Real-capital execution remains gated.