I'm Titus Zhang, a quantitative researcher focused on cross-sectional equities, leakage-aware validation, and cost-aware portfolio construction.
01 / Selected work
Quantitative research · Python
A public, reproducible framework for ranking U.S. equities by forward 10-day returns, with expanding-window validation, purged boundaries, final out-of-sample evidence, and explicit transaction-cost diagnostics.
Research tooling · Local-first
A macOS research workbench that unifies experiment setup, factor diagnostics, backtest review, artifact traceability, and bounded report generation while keeping data and credentials on the user's machine.
Financial data product · Full stack
A trustworthy financial-news inbox that groups related public headlines into stable stories, preserves source lineage, and separates relevance, importance, credibility, and freshness for research triage.
02 / Research practice
Split time before feature construction, purge overlapping labels, and make the signal, execution, and evaluation clocks explicit.
Track turnover, exposure, and transaction-cost sensitivity before interpreting portfolio returns or model rankings.
Separate research evidence from trading claims and keep data, execution, and comparison limits visible beside every result.
03 / Experience
Asymptote Capital · Remote
Evaluated U.S. equity signals, reproduced reference portfolio behavior, and compared model and regime-risk approaches using IC, turnover, costs, drawdown, and portfolio diagnostics.
Oceanum · Shanghai
Built configurable SQL and Python data workflows and extended a credit-policy simulation across approval, profit, staffing, error, and bad-debt trade-offs.
Xinsheng Group · Shanghai
Built an A-share cross-sectional research dataset, benchmarked sequence models against simpler baselines, and evaluated a weekly portfolio with stability, exposure, turnover, and cost controls.
04 / Education
2024 – 2026
M.S. Quantitative Finance & Risk Management · GPA 4.0/4.0
2020 – 2024
B.S. Applied Mathematics