Skip to content
H
Work experience

Jun 2026 — Present

BEI

Financial Modeling Intern · Bitcoin Education Institute

Quantitative research on post-earnings drift and Bitcoin price states, from a C++/CUDA data pipeline to a served analytics dashboard.

QuantTime SeriesLightGBMCUDA

Highlights

  1. 01

    Designed a market-neutral PEAD equity strategy yielding 8.4% annualized return and a 1.15 Sharpe ratio net of transaction costs across 182,712 WRDS/IBES earnings events and 6,766 stocks (1996-2013) via decile-ranked long/short portfolios over 60-day drift windows.

  2. 02

    Structured a defined-risk PEAD options strategy generating 24% average return per trade, a 0.92 Sharpe ratio, and a 57% win rate across 6,766 stocks using vertical call and put spreads on top- and bottom-decile earnings surprises.

  3. 03

    Shipped a React + FastAPI dashboard converting 5,808 daily Bitcoin closes into Bayesian-smoothed probabilities across 80 moving-average/horizon combinations, cutting repeat-analysis latency from 1.2 seconds to under 1 millisecond with TTL-LRU and rolling-mean caches.

  4. 04

    Improved out-of-sample calibration by 35.5 percentage points over a 64.5% base rate, reaching a 100% hit rate across 45 held-out analog days by re-ranking 21,689 parameter combinations with Wilson-score confidence bounds.