• Led team of 5+ data scientists building a revolutionary, deep-learning driven revenue management system for airlines

• Operated as MLE - built machine learning systems which powered an intelligent distributed energy network. Significantly improved forecasting performance.• Built fast simulator which enabled additional revenue sources more powerful and effective sales proposals• Developed analytics system utilized by business teams to make data-driven decisions to pinpoint lost revenue and highlight operational issues

• Evaluated and assessed student progress, providing constructive feedback and support to help them achieve their academic goals

Distributed Matrix Factorization in Spark• Implemented novel matrix factorization algorithm in Spark utilizing stochastic gradient descent • Outperforms built-in ALS implementation with regards to both runtime and MSE for large matrices• Added ability to incorporate item and user side data – not currently possible in Spark MLlibGitHub Recommender• Engineered features from 3+ TB of data to create recommender for GitHub’s OSS projects• Clustered projects based on features including languages used, activity, popularity, etc.Real-Time Fraud Detection App• Trained random forest fraud classifier for online ticket reseller with 91% F1 score • Deployed real-time detection app with Flask

• Slashed costs on $1.2M project by 17% by devising automated data extraction with Python• Modelled stress cracking for a $1M pipeline retrofitting project

• Reduced fines and rig downtime by 15% through innovative well data management system• Generated reservoir simulation studies that underpinned key economic decisions for one of the world’s largest upstream oil producers (market cap over $26B)

• Achieved accurate simulation of FeCr production to pinpoint manufacturing inefficiencies• Established state-of-the-art plasma processing lab through construction of reactor control systems