I make high-stakes conclusions verifiable: certified error bounds in derivative pricing, layered verification in frontier AI evaluation, and evidence standards for delegating authority to autonomous systems.
Duke Alumnus, B.S. in Mathematics and M.Eng. in Risk Engineering · Sequoia Fellow
theodoreoy.com · Email · X
- Critical Dependence Stress for Diversification Gains under Fixed Marginals How much dependence stress reverses a portfolio's Expected Shortfall advantage while every loss distribution stays fixed. In the uniform benchmark, the critical budget is at least 30% above the tangent prediction.
- Certified Valuation of Arithmetic Asian Options What an arithmetic Asian option loses when exercise is decided on the geometric average of its fixings. The proved bound is below a hundredth of a cent per barrel for a one-month crude oil option, and under Heston it becomes a certified price interval.
- Certified Joint Pricing Errors in Rough Heston How far stored option prices can sit from the exact model prices across a whole portfolio, when one transform error is common to every strike. Netting the legs certifies a half-point tolerance for all 28 six-month S&P 500 positions tested, against 13 without.
- Large-Scale Domain Expert Feedback Governance and Adaptive Review A statistical decision framework that prices the measurement and the repair of expert-generated training data in one rule. It coordinated more than 300 specialists, and later post-training programs reused it.
- Financial Task Design and Synthetic Data Engineering A framework for scaling financial training data while preserving the validity of its supervision. I created Batch Zero to test its rules on domain experts before a task family scales, and the partner company adopted the protocol across other post-training programs.
- General Reasoning Evaluation and Layered Verification A proof-obligation framework linking conclusions to their assumptions, derivations and execution evidence. On a test set of 3,600 submissions with known verdicts, the certificate verifier accepted every valid proof and rejected every invalid one.
- Evidence Verification and Robust Comparison in Investment Research A decision-theoretic framework for investment due diligence that determines which conclusions the evidence supports and which facts are most valuable to verify next. I implemented it with a leading global alternative asset manager.
Four articles from Duke Law School on who may decide under risk: autonomous authority in space, small states and strategic space dependence, solar geoengineering, and mangrove restoration and compensatory mitigation.
Jiritsu Network · Euromonitor International · China Construction Bank · Jones Lang LaSalle · Hubble Network · SAIF Partners · CITIC Securities · EY-Parthenon
The site is built with the same care as the work it presents: a Next.js static export, a Canvas 2D clarinet drawn from separated particles, mathematics rendered to SVG at build time, figures drawn from each manuscript's own numbers, legal footnotes that open in place, a strict content security policy, and tests that cover every route and every line of visible copy.