- 🔭 I'm currently working on systematic alpha research with leakage-safe, out-of-sample validation
- 🌱 I'm currently learning robust and stochastic methods for separating persistent signals from unstable backtest results
- 👯 I'm looking to collaborate on quantitative research and reproducible validation workflows
- 💬 Ask me about point-in-time data, walk-forward validation, and alpha attribution
- 📫 Reach me at wl3123@columbia.edu
- ⚡ A failed alpha can still be a successful research result when the failure is explained clearly
A relative-strength rotation study implemented in Python and QuantConnect/LEAN. Factor attribution shows that an apparently strong intercept largely disappears after controlling for the portfolio's direct economic exposures.
An end-to-end US equity research system covering point-in-time reconstruction, leakage-safe walk-forward evaluation, model-risk diagnostics, and tail attribution. The study concludes that no tested ML candidate was stable enough to promote.
The objective is not to make every idea look successful. It is to determine which evidence survives careful validation.
Most repositories here are research and educational projects, not investment advice.