Campaign Entity Resolution
Fuzzy matching + BERT embedding system to deduplicate and resolve campaign entities across CRM and marketing platform data sources.
Akash Sharma
Primary Owner
Overview
Built entity resolution pipeline using Splink for probabilistic record linkage, enhanced with BERT sentence embeddings for semantic matching of campaign names. Resolves campaign variants across systems into canonical entities for accurate attribution and MMX modeling.
The problem
Campaign names across CRM and marketing platforms had inconsistent naming conventions, abbreviations, and typos, preventing accurate join between ad spend data and pipeline impact data.
Results
Enabled clean campaign-to-pipeline joins for attribution and marketing mix modeling by resolving entity mismatch across data sources.
Related Data Engineering Projects
PySpark Data Pipeline Migration
Migrated the revenue metrics pipeline from single-machine pandas to distributed PySpark, resolving critical implementation bugs to unblock daily ML scoring.
Dynamic SQL Query Generation
Metadata-driven SQL generation system translating user-defined filter configurations into executable BigQuery queries.