Summary

Machine Learning Engineer with expertise across industrial and financial domains—building custom gradient boosting algorithms for market prediction and deploying computer vision systems for automated quality control.

PackBoost – Custom Gradient Boosting Algorithm

A domain-specific gradient boosting implementation designed for financial market predictions with constraints beyond standard libraries. Features include:

  • Era-aware splitting for improved robustness across market regimes
  • Synchronized ensemble feature sampling to enforce orthogonality between weak learners
  • Massively parallel training through round-forward sample path reuse

Industrial Machine Vision System

End-to-end computer vision pipeline for automated quality control in steel manufacturing:

  • Real-time weld detection on high-speed production lines
  • Surpassed manual inspection accuracy despite <1% defect rate (extreme class imbalance)
  • On-premise deployment with edge processing for reliability in industrial environments