Interviewing — Full-Time, Fall 2026

Operations management,
built for the real
world of decision-making

PhD candidate in Operations Management at the University of Rochester (M.S. Operations Research, Columbia). My work spans predictive analytics, machine learning, and data-driven decision-making under uncertainty — models rigorous enough to defend and robust enough to ship. Most recently, I delivered a large cancer institute's first production machine-learning model on Databricks, a patient wait-time predictor now live in clinical workflows.

The research-to-impact transition.

My PhD is in Operations Management, but I've spent it on the kinds of problems industry faces everywhere: time-series forecasting, calibrated probabilistic prediction, and stochastic optimization under uncertainty. These questions reached me through a healthcare partnership — the methods travel far beyond it.

Each of my three research projects goes all the way from theory to a shipped tool. I've worked with over a million large-scale records, built simulation platforms from scratch, and turned predictive pipelines into human-in-the-loop systems — one deployed as a large cancer institute's first production ML model on Databricks, another as a Python-plus-Excel interface that operations staff actually open.

That instinct traces back to my pre-PhD work as a quantitative researcher in finance: backtesting trading strategies, calibrating Heston stochastic volatility models, and deriving option-implied probability distributions on the S&P 500. I'm seeking full-time roles (quantitative researcher, data scientist, or consultant), where I can keep pairing rigorous modeling with messy real-world decisions.

Featured Projects

2023  —  2026
Project 01 · Probabilistic ML
Paper Under Review

Dynamic Interval Prediction for Patient Wait Times

Reframed patient wait times as a probabilistic-interval problem rather than a point estimate. Adaptive quantile selection on Random Survival Forests / a pinball-loss neural network, with real-time conditional updater and SHAP explainability. Deployed in production at a large cancer center — now serving live predictions to clinical assistants in five-minute refresh cycles. 79% accuracy, 38-min interval width.

Random Survival Forests Neural Networks SHAP PyTorch
Project 02 · Time-Series Forecasting

Probabilistic Forecasting of Intraday Resource Utilization

A set function learning problem — the input is a set of scheduled sessions, the output is the intraday occupancy curve. Three complementary routes: (1) top-down / bottom-up ML to forecast daily totals, then shape with Beta or Gaussian-process fits; (2) direct aggregation of Monte Carlo session simulations; (3) PCA on historical curves with ARIMAX. Shipped behind an Excel interface that lets planners edit sessions and see the load update live.

Set Function Learning Monte Carlo XGBoost · ANN Gaussian Processes ARIMAX · PCA Python + VBA
Project 03 · Simulation + Optimization

Can We Predict the Day Ahead? A Study of Appointment-Based Healthcare Service Systems

A simulation-optimization framework that predicts how a day will unfold given an appointment schedule — and quantifies the irreducible error of the underlying stochastic service system. +5–28% accuracy over heuristic baselines.

Simulation-Optimization Discrete-Event Sim Queueing Theory Stochastic Modeling Python

Work Experience

Industry positions where the work had to ship, not just publish.

May — Aug 2025

Data Scientist Intern

Dana-Farber Cancer Institute · Boston, MA
  • Designed and deployed the institute's first production ML model on Databricks — a patient wait-time predictor with 80% accuracy, reducing manual estimation time by ~5 min per patient.
  • Engineered scalable pipelines (8,200+ lines of code, 3,500+ lines of tests) over 1M+ patient records, with automated retraining and drift monitoring.
  • Collaborated with 15+ clinicians, operations managers, and engineers to translate business requirements into analytical models and ensure adoption.
Sep 2020 — Apr 2021

Quantitative Researcher

Veta Investment Partners · Topeka, KS
  • Developed and backtested SPY trading algorithms using volatility and momentum signals in Python, achieving a 70% cumulative improvement over benchmarks.
  • Built a quantitative framework for option valuation and ETF ranking with scenario analysis and stress testing.
  • Automated intraday portfolio attribution and risk reporting by extracting data from SQL, computing performance metrics, and delivering real-time email notifications.
Feb 2020 — Sep 2020

Quantitative Researcher

Wisdom Capital Asset Management · New York, NY
  • Derived option-implied probability distributions of the S&P 500 using non-parametric methods, calibrated with the Heston stochastic volatility model to improve predictive accuracy.
  • Conducted risk analysis and scenario testing across short equity-hedged option portfolios, incorporating implied volatility, stress testing, and Value at Risk (VaR).
  • Analyzed 20+ assets across 30 years of return distributions.

Education

A foundation in operations research, applied statistics, and financial engineering.

2021 — 2026

Ph.D., Operations Management

Simon Business School, University of Rochester · GPA 3.9/4.0

Three concurrent research projects on healthcare operations — probabilistic ML, forecasting, and stochastic scheduling. Coursework in Causal Inference, Algorithm Design, and Probability & Statistics.

2018 — 2020

M.S., Operations Research

Columbia University · GPA 3.9/4.0

Coursework in Machine Learning, Data Analytics, Simulation, Applications Programming, and Tools for Analytics.

2014 — 2018

B.A., Financial Engineering

Wuhan University · GPA 3.8/4.0

Quantitative finance, stochastic processes, and statistical modeling.

Invited Talks

Oct 2025 Dynamic Interval Prediction for Patient Wait Times in a Cancer-Care Facility INFORMS Annual · Atlanta
Jul 2024 Dynamic Interval Prediction for Patient Wait Times in a Cancer-Care Facility INFORMS MSOM · Minneapolis
Aug 2023 Physician Rostering Problem with Downstream Capacity Constraints INFORMS Healthcare · Toronto

Technical proficiencies.

The tools and methods I reach for across research projects and production work.

01Modeling
Predictive models · Time-series models · Monte-Carlo simulation · Optimization · Queueing theory · Scenario analysis & stress testing
02Machine Learning
Random survival forests · XGBoost · Quantile regression · Neural networks · SHAP explainability · Probabilistic calibration
03Engineering
Python · PyTorch · scikit-learn · Git · Unit testing · MLOps pipelines · VBA
04Data & Cloud
Databricks · PySpark · SQL · Azure · MLflow · CI / CD · Large-scale data pipelines
05Domain
Forecasting & decision systems · Real-time prediction pipelines · Resource allocation under uncertainty · Queueing & service operations · Trading strategies · Option pricing & risk analytics · Backtesting
06Communication
English (fluent) · Mandarin (native) · Technical writing · Cross-functional stakeholder communication · Conference presentation

Let's discuss how my research can impact your team.

Actively interviewing for Quantitative Researcher, Data Scientist, Applied Scientist, and Operations Research Scientist roles for Fall 2026. Happy to chat about any of the work above — or a problem you're trying to solve.