Looking for a job

kevin ross

@kevinross2

Message

Senior Data Scientist at DoorDash measuring LLM-powered discovery and building marketplace experiments.

United States

Message

At DoorDash, I define evaluation and online measurement for LLM-powered discovery experiences, including conversational search and grounded store and item recommendations. I build evaluation datasets, compare LLM-as-judge scores with human review, and design experiments with GenAI-specific guardrails.

Previously at DoorDash, Best Buy, and eBay, I developed models and experiments for marketplace incentives, personalization, recommendations, and customer behavior. My work includes uplift modeling, causal evaluation, and measurement frameworks that connect product changes to business outcomes.

Experience

Work history, roles, and key accomplishments

Current

Senior Data Scientist

Current

DoorDash

Sep 2021 - Present (5 years 1 month)

Defines evaluation and online measurement strategy for LLM-powered discovery experiences, including conversational search and grounded recommendations. Developed uplift models for promotion targeting and applied quasi-experimental methods to evaluate marketplace changes.

LLM Evaluation LLM As Judge A B Testing CUPED Switchback Testing Causal Inference Uplift Modeling Individualized Treatment Effects Experimentation Standards Power Analysis Sequential Testing Metric Frameworks Python SQL Spark Databricks Snowflake BigQuery

Data Scientist

Best Buy

Jul 2018 - Sep 2021 (3 years 2 months)

Developed customer-level features and propensity models for marketing analytics and personalization. Built recommendation and ranking models, designed experiments, and supported migration to GCP and BigQuery.

Python SQL R Spark BigQuery GCP Propensity Modeling Recommendation Systems Personalization A B Testing Causal Inference Demand Forecasting Feature Engineering Marketing Analytics

Data Scientist

eBay

Oct 2016 - Jun 2018 (1 year 8 months)

Analyzed buyer discovery funnel and developed predictive models for search and recommendation. Built customer segmentation and analyzed A/B experiments for product changes.