Paper-first gates
Research and execution paths remain separated by explicit authorization, evidence, and lifecycle gates.
Research engineering
Trading systems treated as evidence problems.
These are personal research systems used to practice data validation, reproducibility, lifecycle correctness, auditability, and risk controls. They are not investment products for sale, and this site makes no claim of demonstrated edge, profitability, annual volume, or win rate.
01Implemented capabilities
What the systems demonstrate technically
The useful evidence is engineering discipline, not a financial-performance story.
Reproducible records
Deterministic reports, manifest hashes, cost-aware comparisons, and accounting records designed for independent verification.
Fail closed on ambiguity
Malformed or contradictory market data blocks progress instead of being silently reinterpreted to make a test pass.
Intent before action
Order intent, uncertain placement, reconciliation, restarts, and shutdown behavior are treated as state-machine problems.
Runtime truth
Process state, deployed bytes, audit events, exposure, and infrastructure are checked directly rather than inferred from configuration files.
Negative results count
A system that has not demonstrated an edge is recorded that way. Activity or a favorable short run is not treated as proof.
Discuss the system or the engineering behind it.
For a technical role, implementation engagement, or focused engineering discussion, contact Sean directly and identify the research that brought you here.