I'm a UW鈥揗adison student building software to explore markets, uncertainty, and decision-making. My projects span Python/C++ systems, numerical research, full-stack applications, and bounded AI workflows.
Seeking Summer 2027 internships 路 Software engineering 路 Quant development 路 Fintech
Python C++17 TypeScript / React SQLite
A research application that turns supported market-microstructure questions into validated simulation plans, runs experiments across seeds, and produces reports linked to evidence.
- Bounded orchestration with work budgets, cancellation, checkpoint recovery, and transactional run history.
- Optional AI planning and review; Python controls execution, numerical results, and report validation.
- An underlying price-time matching engine with cancel/replace rules, Python/C++ parity tests, and repeated API benchmarks.
The public demo uses scripted planning and real simulations. Live-model integration is implemented with mocked-transport tests; live behavior and any multi-agent advantage remain unmeasured.
Try the demo 路 Workflow, evaluations & recovery
The free demo may take a minute to wake up.
Python NumPy SciPy
How much pricing error remains for a given compute budget? Compare Monte Carlo, antithetic sampling, control variates, scrambled Sobol, and importance sampling using reproducible synthetic contracts. Measure price and delta error, runtime, and rare-payoff failure cases against analytical references.
Study design & reproducibility
Python pandas SQLite
An event-driven portfolio simulator with next-open execution, explicit costs, cash reservation, and auditable ledgers. Stored configurations and price inputs support exact replay; a separate selection-bias study examines how searching many strategies can manufacture an impressive in-sample winner.
Python TypeScript
CFR/CFR+ for decisions under hidden information, with exact best-response and exploitability evaluation. Includes resumable training experiments and an interactive explorer for inspecting action probabilities, regrets, and playing against a saved policy.
TypeScript Next.js Prisma PostgreSQL
A paper-trading journal for stocks, options, and futures, with portfolio views, market-data integrations, and optional Gemini educational commentary. An application demo; authentication and user isolation are still future work.
| Project | Focus |
|---|---|
| Momentum Factor Research | Monthly momentum experiments with explicit costs, missing-data coverage, leakage tests, block-bootstrap inference, and SQLite provenance. |
| New Space Radar | Source-aware event research for space equities, exchange-calendar alignment, synthetic validation, and a TypeScript evidence explorer. |
| Leveraged ETF Risk Lab | Portfolio accounting, risk attribution, and daily-reset leverage, with hand-calculated fixtures and reproducible offline reports. |
- Make results reproducible: preserve configurations, seeds, inputs, and evidence.
- Test the failure cases: invalid plans, interrupted runs, timing errors, and misleading estimates.
- Keep claims tied to measurements: distinguish synthetic examples, historical research, and live-model evaluations.
Tools: Python 路 C++ 路 TypeScript 路 React / Next.js 路 SQL 路 SQLite 路 PostgreSQL 路 NumPy 路 SciPy 路 pandas