I keep returning to the seam between the paper version of a thing and the way it behaves in the world: what people do outside the model, where money moves inside an institution, and who finally makes the decision. Behavioral science and climate finance look far apart. To me they ask for the same discipline: don't take a system at its own description. Watch how it works, then decide what to believe.
The course gave us redacted documents from a real wind project. My individual model joined generation, operating cash flow, debt constraints, and equity returns into one chain of consequence. It could show how stress moved through the structure. It could not decide how much risk a person should accept.
A model is most useful—and most honest—when it lays the judgment open and shows where each assumption sends the future.
Continue reading →For four months, our six-person team asked where quantum technologies might actually matter to the energy industry—and how soon. A common three-part framework let us compare six applications without pretending that technical promise and commercial readiness were the same thing.
A set of individual assignments built from a redacted wind-project case: contract reviews, tax exercises, and a financial model linking operating cash flow, PTCs, sculpted debt, and P50/P99 cases.
I supported Lisa Sachs's graduate course on how climate finance is built and constrained: blended and concessional capital, sovereign debt, carbon markets, the UNFCCC system, and the cost of capital in emerging markets.
An early research prototype testing whether PPA terms can be extracted reliably from public documents. Model validation, IRA tax-equity review, and cited investment-memo drafting belong to the hoped-for V1—not the product that exists today.
A small tool that rewrites only the bullets in a LaTeX résumé for a particular job—and refuses to invent a better past. I built it after spending too many recruiting-season evenings making the same careful edits by hand.
A writing workspace that begins with notes rather than a transcript. It helps turn meetings, fragments, and audio into structured bilingual documents while preserving the memory of a project across time.
I spent four undergraduate years studying behavioral science. It did not give me a formula for people. It taught me to face a messy situation, resist the quick explanation, and ask: what is actually happening here, and how could we know?
Project finance turned out to reward the same habit. A model appears to calculate, but it is always making an argument about the future; the consequential claims are often tucked into the quietest assumptions.