ABACUS — Advanced Credit Risk Prediction Analysis Tool

Company: AZEN Global, South Korea · Role: Data Scientist / ML Engineer

ABACUS is a platform that leverages machine learning to enhance fault detection capabilities and refine credit risk modeling, significantly improving the precision of financial assessments and operational reliability for over 10 million users.

Key Contributions

• Developed advanced credit risk models by implementing machine learning algorithms, achieving a 3% lift in credit risk scoring accuracy and sharpening the precision of downstream financial assessments.
• Optimized data pipelines for fault-detection systems by streamlining processing on Apache Spark, materially increasing throughput for workflows serving 10M+ users.
• Architected highly-available, resilient database clusters across ClickHouse · MongoDB · Redis to manage 5B+ records, with disaster-recovery provisions safeguarding data integrity and operational continuity.
• Built real-time monitoring dashboards and interactive reports surfacing live platform signals, strengthening operational transparency and supporting informed decision-making.

Tech Stack

Python · Django · FastAPI · ClickHouse · MongoDB · Oracle · Redis · Apache Spark