AI research
Probabilistic outputs. Evidence, hypotheses and uncertainty; no authority to rewrite limits.
INVESTMENT APPROACH
A thesis is the beginning. It has to earn the rest.
The system we are building connects nine stages, with a visible record of the evidence and authority behind each decision.
- 01
Data
Trace observations to their source. Check timestamps, coverage and permissions.
- 02
Research
Look for relationships that can be explained, challenged and tested.
- 03
Thesis
State the hypothesis, horizon, evidence and conditions that would invalidate it.
- 04
Debate
Ask independent reviewers to identify missing evidence and competing explanations.
- 05
Portfolio
Translate a surviving idea into a proposed allocation with explicit constraints.
- 06
Risk
Apply deterministic eligibility, sizing, concentration and execution limits.
- 07
Execution
Route only authorized orders through brokerage infrastructure and reconcile results.
- 08
Monitoring
Observe positions, data freshness, system health and exceptions.
- 09
Evaluation
Compare outcomes with a frozen baseline, costs and the original hypothesis.
SEPARATE RESPONSIBILITIES
Every layer has a job.
Risk controls
Deterministic checks. Eligibility, sizing and execution constraints are independently enforced.
Human governance
Mandates, model approval, exceptions and emergency intervention remain human responsibilities.
Brokerage
Authorized order routing and account records. A connected broker account is not a pooled fund subscription.
Fund administration
Independent accounting and official investor reporting are a separate integration. No administrator appointment is represented at this stage.