Quantitative research infrastructure for people who want to do this properly.
Small group. Factor models, regime detection, risk decomposition. We build the backtesting and analysis tools ourselves. Everything is open-source.
Research
Factor work
We study what drives returns, how those drivers decay, and when they stop working. Most of our current time goes into finding reasons to reject hypotheses.
Infrastructure
Kernel
Research infrastructure. Right now: clean data and sec master. Planned: backtest engine, factor framework, risk and regime tooling. See Kernel.
Process
Risk-first
We think about what can go wrong before we think about what might go right. Capital preservation over returns.
Active research
What we're working on
- Covariance · Regime How should covariance matrices be regularized to remain stable under regime transitions?
- Correlation · Diversification Does correlation structure contain early information about diversification decay?
Open problems
We maintain a list of research and engineering problems. Pick one, write a proposal, and if the work is good we'll reach out.
View problems