Anomaly Detection System

Context: Global commodity pricing platform where data integrity across development, testing, and production environments was critical.

Challenge: Record mismatches between environments – including prices, users, traders, and deals – required manual oversight and were prone to human error.

Solution: Designed and implemented an ML‑based anomaly detection system that:

• Continuously monitored record mismatches across dev/test/prod environments
• Automatically flagged inconsistencies in price trends, user activity, and deal records
• Integrated directly with existing R Shiny dashboards for real‑time visibility

Outcome: Drastically reduced manual oversight, improved data trust, and ensured that editorial teams always worked with accurate, consistent data across all environments.