Approach

Portfolio construction under real constraints: placing assets across accounts with different tax treatment, keeping index portfolios within tolerance of their benchmarks, and translating investment methodology into orders a portfolio manager can approve.

Decisions whose value appears over long horizons or in the tails, where the object of interest is a distribution across simulated paths rather than a point estimate.

Formulate first. Decision variables, objective, and constraints come before implementation; in portfolio problems most of the judgment sits in the constraints.

Evaluate under uncertainty. Judge a policy by simulating it against a baseline over the relevant horizon, not by a single-period objective value.

State the model scope. Every result is conditional on its assumptions, and the assumptions that drive it are part of the result.

Summer 2027 internships in quantitative research, portfolio management, and systematic investing, during the UC Berkeley Haas MFE.