Overview
Architected a scalable database and ETL pipeline for a multilingual learning platform.
Role
Data Engineer
Project Details
Engineered a production ETL pipeline migrating 24K+ entities into a normalized PostgreSQL schema, generating 985K+ language exercises with <2% failure rate.
Built an LLM-powered linguistic classification pipeline (Google Gemini) with JSON validation to dynamically tag and categorize language complexity metrics.
Optimized end-to-end migration runtime from 12+ hours to 45 minutes by implementing B-tree indexing strategies, optimized queries, and parallel batch execution.
Implemented scalable JSONB metadata architecture to support Extensible schemas without relational migrations.