University of California, Berkeley

Software Development Engineer II at Amazon

San Francisco Bay Area

Forecasting Deep Learning
-  Working on model improvements for demand forecasting to improve forecast accuracy.
-  Developed a novel model architecture for multi-horizon forecasting which saw significant accuracy improvements against SOTA architectures.

Amazon

Software Development Engineer II, SCOT Forecasting

Forecasting Deep Learning
-  Working on large scale ML systems for time series demand forecasting to forecast for millions of products daily.
-  Worked on distributed training as a means to scale training to improve model accuracy on a wider distribution of products.
-  AutoML: Worked on automation of model iterations (data pre-processing, training, backtesting).

Also help teach ML University:
-  Math for Machine Learning
-  Decision Trees and Ensemble Methods

Software Development Engineer I, SCOT Forecasting

Forecasting Deep Learning
-  Worked on Distributed Training

Yagna Patel