Experience snapshot
Censia
Marketplace analytics & forecasting
Data Engineer.
Built forecasting, KPI, and experimentation infrastructure that turned fragmented marketplace data into a clearer operating picture for supply-and-demand decisions.
The context.
Problem space
Marketplace data was fragmented across systems, making it difficult to understand listing performance, prioritize supply and demand, and compare product initiatives consistently.
Responsibility
Own forecasting and analytics pipelines, consolidate KPI data in SQL and BigQuery, and define a repeatable experimentation standard for product and model work.
Outcome
Improved forecast quality, accelerated analysis, and gave stakeholders a more reliable framework for prioritization and experiment readouts.
What I owned.
01
Built forecasting pipelines end to end in Python and PyTorch to identify the drivers of listing performance and support marketplace prioritization.
02
Consolidated fragmented hiring and marketplace data into analyst-ready KPI datasets in SQL and BigQuery, creating a reliable reporting base for the business.
03
Defined the experimentation standard across eight product and model initiatives, including success metrics, guardrails, and a consistent readout format.
04
Used forecasting and marketplace analysis to help the business decide where to prioritize limited supply-and-demand resources.
05
Presented KPI trends, experiment results, and tradeoffs directly to business stakeholders so technical findings translated into operating decisions.
The evidence.
22%
reduction in prediction error from the forecasting pipeline
35%
faster query performance after rebuilding KPI datasets
8
product and model initiatives using the experimentation standard
A note on confidentiality.
This experience involved internal marketplace data and operating decisions. The overview focuses on my responsibilities, methods, and résumé-level outcomes; customer information, proprietary datasets, and internal implementation details are omitted.
PORTFOLIOBack to home
↗