Looking for a job

Cole Liu

@collinlyou

Message

I build experimentation, causal inference, and predictive models that improve consumer-product decisions and outcomes.

United States

Message

What I'm looking for

I'm looking to lead decision-grade experimentation, causal analysis, and predictive modeling for consumer products, partnering with ML engineering to bring models into production and mentoring teams on statistical rigor.

I've improved consumer-product decisions at TikTok, Tinder, Snap, Headspace, and Thrive Market through experimentation, causal analysis, forecasting, and predictive modeling. At TikTok Shop, I built an eight-stage buyer-funnel metric layer, established experiment guardrails across U.S. Shop pods, and reduced median A/B test duration from about 21 to 13 days.

I've moved models into production with ML engineering teams, including seller-retention propensity scoring at TikTok and LTV scoring at Tinder. My work spans retention, churn, LTV, pricing, onboarding, and subscription funnels, and I mentor data scientists on statistical rigor.

Experience

Work history, roles, and key accomplishments

Current

Senior Data Scientist

Current

TikTok

Jul 2024 - Present (2 years 2 months)

Designed TikTok Shop's 8-stage buyer-funnel metric layer in SQL on Hive/Presto, authored the Shop-wide experiment guardrail framework, and implemented CUPED variance reduction. Built survival models and partnered with ML engineering to operationalize a seller-retention propensity score.

SQL Hive Presto Python Statsmodels CUPED Kaplan Meier Cox Proportional Hazards LIFELINE PyTorch Causal Inference Experimentation

Senior Data Scientist

Tinder

Jan 2023 - Jul 2024 (1 year 6 months)

Achieved ~14% WAPE on a 90-day LightGBM LTV model, re-architected the payer LTV scoring workflow into a PySpark pipeline, and designed a geo-staggered weekly-plan experiment. Developed a deep neural network engagement model validated through A/B testing.

LightGBM PySpark Databricks MLFlow Docker Python A B Testing LTV Modeling deep neural networks

Data Scientist

Snap Inc.

Sep 2019 - Sep 2022 (3 years)

Shipped 7 of ~40 A/B tests with day-7 retention lifts on Snapchat onboarding and Stories, launched engagement-parity monitoring for the Android rebuild, and modeled day-7 retention with logistic regression. Developed a deep neural network engagement-prediction model.

A B Testing Looker BigQuery scikit learn SHAP Logistic Regression deep neural networks Python

Data Scientist

Headspace Inc.

Dec 2018 - Sep 2019 (9 months)

Conducted 11 A/B tests on the Headspace app trial-to-paid funnel, trained an XGBoost churn model, and produced a rolling paid-subscriber forecast. Authored the Airflow DAG automating weekly scoring to CRM.

A B Testing XGBoost Airflow Python Churn Modeling Forecasting

Product Analytics Engineer

Thrive Market

Feb 2017 - Dec 2018 (1 year 10 months)

Established the product-analytics foundation by authoring a unified ~120-event tracking schema, replaced manual Excel reporting with a membership funnel in Redshift SQL and LookML, and engineered checkout A/B readouts.

SQL Redshift LookML Optimizely Python A B Testing Funnel Analysis