Levi Strauss & Co.Machine Learning Engineer
Aug. 2020 - Dec. 2021Remote• Implemented Dockerized LSTM Deep Learning Classification Models as REST APIs within a Google Kubernetes Engine Cluster. • Revamped initial model iterations, resolving inconsistencies and low-quality product recommendations (complementary, similar, personalized). Architected LSTM model, integrated Keras-Tuning, Tensorboard, and MLOps principles, resulting in a 2x improvement in metrics. • Implemented Dockerized XGBoost models as Machine Learning Pipelines utilizing Apache Airflow and AWS Batch to address Adjusted Demand and Out-Of-Stock issues within the supply chain. Elevated the process from manual deployments and overrides to a production-ready application through the leadership of an engineering team, resulting in cost savings for the company. • Developed a company-wide Machine Learning Operations Architecture to standardize infrastructure development, model output versioning, feature handling, and integration into a centralized ML model registry supported by MLFlow. This initiative was crucial as our Machine Learning Engineering team expanded, ensuring consistency and efficiency in our processes.