CombinedMachine Learning and Continuum-scale Modeling to Predict the Behavior of Li Ion Batteries

Maintained ETL Pipelines, builtData Warehouses, and productionized Data Science workflows using microservices on kubernetes

Scoped and lead development of in-house predictive analytics platform and data warehouse for battery data

Improving the process of fitting electrochemical models with inspiration from machine-learning-based Chess Engines. Reduced in-house Li-Ion model calibration time by 60% and improved fit by 30x over Genetic Algorithm

Publication

Generated Ensembled Neural Networks that can accurately represent internal battery states while enforcing physical constraints, including monotonicity. Sped up implementation 5x leveraging JIT-compiled functions on NumPy arrays.

Application

A platform for learning music theory and building your piano repertoire leveraging advanced analytics and spaced repetition.

Application