Alexander Rom | Portfolio

About Me

I'm a quantitative researcher and trader with a background in econometrics and computer science, completing my M.S. in Financial Engineering at USC. I work across volatility, derivatives, and macro, developing models and building the systems needed to turn research into trading decisions. My current research interests center on connecting probability distributions implied by derivatives prices with real-world processes and outcomes and exploring what alternative data, including prediction markets, reveal about expectations, volatility, and tail risk, especially when regimes shift and familiar assumptions break down. Additionally, I build structured research pipelines that use LLMs to derive and combine methods from the literature, test new hypotheses against market data, and preserve experimental results in a growing knowledge base that guides and accelerates subsequent research. This site brings together my research and ongoing work.

Research & Projects

Deriving a Markov Transition Model from Option-Implied Risk-Neutral Densities

Details →

Options · Risk-Neutral Density · Markov Chains · Derivatives Pricing

Constructs a discrete-time Markov chain consistent with option-implied marginal distributions, enforcing risk-neutral martingale constraints via local tri-diagonal transitions.

Open Source Python Library for Kalshi Prediction Market Data

Details →

Prediction Markets · Market Data · Time Series · Python Library · Open Source

Open source Python library that pulls candles, trades, and order books from Kalshi prediction markets into structured data. Real collected data example included, from years of daily prices down to individual fills.

Mapping Market Stress Concentration via Put–Call Ratios

Details →

Derivatives · Options Microstructure · Market Structure · Python

Builds signal-generating diagnostic tooling to highlight where market stress and downside hedging pressure may be concentrating, by decomposing index option activity into flow, positioning, and term-structure components across expirations.

Time-Series Analysis of Treasury Bond ETF

Details →

Time Series · ARIMAX · TAR Regimes · GARCH · Forecasting

Modeled daily TLT returns (2004–2024) with regime-switching ARIMAX (Fed-rate states) and GARCH volatility, showing strong flight-to-safety dynamics (VIX↑, equities↓) and improved fit/forecasting versus a baseline ARIMA.

Impact of Banking Competition on Household Loan Rates in the Euro Area

Details →

Panel Data · Hausman–Taylor · Fixed Effects · Banking / Credit Markets

Panel regression across 13 Euro-area countries (2014–2020) testing how local banking competition (branches per 100k adults) relates to household loan rates using Fixed Effects and Hausman–Taylor.

Determinants of U.S. Credit Card Delinquency Rates

Details →

Time Series · OLS Regression · Nonlinear Effects · Newey–West · Household Finance

Models U.S. credit-card delinquency (2000Q3–2023Q2) using quarterly FRED data and an OLS/Newey–West time-series regression to quantify how credit-card APRs, financial conditions, and unemployment (with a COVID regime interaction) explain default risk over time.

Coursework & Notes

Machine Learning for Data Science

Details →

Graduate Coursework · 8 Projects · 189-Page Notes · Python

Eight machine learning projects, from k-nearest neighbours to transfer learning with pretrained networks, paired with a 189-page typeset notes book covering the full course.

Financial Engineering

Details →

Graduate Coursework · 5 Projects · 196-Page Notes · Python

Five quantitative finance projects across rates, volatility, exotic derivatives, and credit, from curve bootstrapping to a 30-year guarantee valuation, paired with a 196-page course notes book.

Experience

Savoia Investments

Quantitative Researcher Intern

San Francisco, CA May 2026 – Aug 2026

I designed, built, and shipped five production systems in Python for a derivatives-focused hedge fund in one summer: a volatility forecasting engine, an options screener, a research dashboard, a broker execution layer, and a daily portfolio risk program. All five run daily on the live desk, on real capital, backed by 1,229 passing test cases.
Volatility engine: 17 realized volatility measures and 15 conditional volatility models, every measure-model combination tuned across inputs, training windows, and scoring metrics in a leakage-proof walk-forward design. A pre-registered study across four horizons ranks each candidate against its peers and a baseline, and the winning models ship with calibrated prediction intervals and live drift monitoring.
Screener: fast, multi-condition scans across the full US-listed options market, gating on listing structure, pricing, and liquidity, then convexity analytics computed over each contract's own greek and volatility history, so present pricing ranks against its own past. Shortlisted candidates reach the dashboard I built on the broker library, where they are repriced live and judged against the book they would enter before the trade is placed.
Portfolio risk and analysis: full book measurement resting on maturity-corrections derived by me, which normalize greek time series across tenor, so an extreme rank signals a regime change, not the passage of time. Any sub-portfolio or strategy isolates across 75 daily metrics and 48 history statistics, driving exit and hedge decisions on live positions, with every number queryable in plain SQL and any past day reproducible byte for byte.

Insider Ownership Index

Quantitative Research Consultant

Los Angeles, CA Oct 2025 – Nov 2025

Derived a return-maximizing insider ownership curve for S&P 500 firms using cross-sectional and panel regressions on 5- and 7-year forward returns; the relationship held across specifications, and the curve captures the marginal effect of insider ownership at each level.
Performed stress testing and scenario analysis on the index with optimal insider-ownership weights to evaluate return stability across market regimes, and estimated Fama–French factor exposures to separate the index's returns from known risk factors.

University of San Francisco

Teaching Assistant

San Francisco, CA Jan 2024 – May 2025

Assisted professors across three courses (Economic Methods, Intermediate Microeconomics, Applied Econometrics); led weekly sections and office hours covering microeconomic theory, regression analysis, and optimization for 60+ students. Across semesters, average course performance increased by ~5% relative to the previous term. In Applied Econometrics, mentored 30 students through thesis-style empirical papers, supporting research design, model specification, sensitivity checks, and publication-ready writing.

Bridges and Barriers Advisory Services

Data Analyst Intern

San Francisco, CA Jun 2024 – Aug 2024

Built Python pipelines for a hedge fund client that ingest, clean, and reconcile market and fundamental datasets into a single production-ready dataset, and automated trade-ops reporting from raw executions to formatted trade tickets, cutting roughly 50 hours of manual work per month and about $5.6K per month in operating costs.
Worked directly with clients to map their operational workflows, identify where manual steps were costing time and accuracy, and propose the automations that addressed them.

Education

University of Southern California (Viterbi)

Aug 2025 – May 2027 (Expected)

M.S. Financial Engineering

Relevant coursework: Probability Theory, Stochastic Processes & Ito Calculus, Machine Learning, Monte Carlo Methods, Derivatives Pricing & Hedging, Volatility Modeling, Interest Rate & Credit Risk Models, Corporate Finance, Market Design & Auction Theory

Activities: Trojan Chess Club

Dean's Master's Scholarship (merit-based)

University of San Francisco

Aug 2022 – May 2025

B.S. Economics (Financial Economics)
Minor: Computer Science

Relevant coursework: Real Analysis, Linear Algebra & Probability, Constrained Optimization, Statistics, Financial & Applied Econometrics, Micro & Macroeconomics, Data Structures & Algorithms, Options & Futures

Fed Challenge (2023): Led 5-person team; presented policy recommendation to Federal Reserve judges