Approach
Start AI-first, instead of adding AI to a team of people.
Most human fund managers lose to the market they are paid to beat. The causes are structural, and a better analyst does not remove them.
The problem
Three limits that people cannot engineer away.
- BiasA losing position becomes a personal conviction.
- LimitsNo team can read every filing and every call.
- MemoryWhen a manager leaves, what they learned leaves too.
Our answer
An investment process designed around AI from the first line of code.
Language models do the reading and the research at a scale no analyst team can match. Every result is recorded, so nothing is forgotten and each new idea starts from everything already tested.
The models themselves improve every few months and are available to anyone. The lasting work is governing them: a clear mandate, hard controls, and a permanent record of what worked and what did not.
Structure
Many strategies, not one or two.
The system is multi-strategy by design. Each strategy has its own thesis and risk profile, and capital allocation across them is designed to follow the evidence, moving towards what works and away from what does not.
Every strategy is recorded through its full life, from first idea to retirement.