Quantitative researcher and research engineer focused on derivatives, volatility, fixed income, and statistical arbitrage.
I build and audit empirical research systems. My work emphasizes point-in-time data, executable timing, realistic costs, leakage controls, multiple-testing discipline, and reproducible metrics. When a result does not survive review, I publish the failure and correct the claim.
I am the founder and quantitative researcher at Navnoor Bawa Research LLC, a freelance markets analyst at FX Empire, and a Computer Science and Engineering undergraduate at Thapar Institute of Engineering & Technology (May 2027). Previously, I worked at Quant Insider and as a WorldQuant BRAIN Research Consultant.
Research writing | LinkedIn | Medium | YouTube | Email
| Program | Validation record | Current conclusion |
|---|---|---|
| Russell 3000 Statistical Arbitrage (site) | 19 walk-forward windows; only W10-W19 are selection-clean out-of-sample (OOS). Five cost profiles. Survivorship bias is disclosed. | No deployable alpha under next-day execution (p = 0.83). The null result is retained. |
| Joint SPX/VIX Calibration and Volatility System (live) | 637 tests. Four audit corrections covering look-ahead, a circular feature, label noise, and strike rolling. | SPX smile RMSE: 0.52 vol points. The VIX-options leg is disabled for structural misspecification; P&L remains model-implied. |
| WTI Crude Oil Volatility Research (live) | 439 purged OOS weeks. HAR-IV compared with mean-reversion and persistence baselines. | Volatility direction: 72.7% vs 65.1%. Level R-squared: 0.50 vs 0.32. The leaked direction strategy was retracted. |
| U.S. Treasury Rates Monitor (live) | Direct Treasury XML, Federal Reserve H.15 history, and deterministic release checks. | Official daily CMT analytics and delayed CBOT futures proxies remain separate by design. |
Additional derivatives work: SABR interest-rate volatility smile engine | options pricing and strategy analyzer
Winner, Trilemma Beta Global Data Science Tournament (2024) - probabilistic Bitcoin-return modeling in a field of 214 teams.
I treat the audit trail as part of the result. Every public claim should state its information set, execution timestamp, sample construction, missing-data treatment, costs, baselines, validation protocol, and known failure modes. I prefer a reproducible negative result to a profitable backtest that depends on hidden assumptions.
- Research: time-series analysis, statistical inference, walk-forward validation, factor models, volatility modeling, derivatives pricing, Monte Carlo simulation
- Engineering: Python, C++, SQL, NumPy, pandas, SciPy, PyTorch, scikit-learn, XGBoost, QuantLib, TypeScript, React, Linux, Git, CI
Open to quantitative research, derivatives, fixed-income analytics, and research engineering opportunities.