Software

Tools

Some software tools that I have been a primary contributor to include:

• ptutils, a set of utilities for training and validating PyTorch models on GPU and TPU.
• tnn, a set of utilities for building temporal neural networks with TensorFlow.
• tfutils, a set of utilities for training and validating TensorFlow models on GPU and TPU.
• Co-author of GRUV, a package for algorithmic music generation using recurrent neural networks.
• Author of Poisson loss, Permute, UpSample1D, and UpSample2D layers in the Keras deep learning API.
• Author of (now defunct) keras-extra package to connect CNN layers with RNN layers in the Keras deep learning API. Initial release: 2015.
• Author of ranking model for Total Waterpolo, which ranks teams based on strength of schedule and home team advantage. Initial release: 2011.
• Author of DMG Automounter for Linux. 19,138 downloads as of September 2021. Initial release: 2008.

Course Projects

Some course projects done during my undergraduate and master’s:

• Recurrent versus Recursive Approaches Towards Compositionality in Semantic Vector Spaces (CS 224U: Natural Language Understanding, Spring Quarter 2016) [code]

• CARMA: A Deep Reinforcement Learning Approach to Autonomous Driving (CS 239: Advanced Topics in Sequential Decision Making, Winter Quarter 2016)

• Inferring Network Dynamics from Simulated Extracellular Electrophysiological Data (CS 379C: Computational Models of Neocortex, Spring Quarter 2015) [code]

• GRUV: Algorithmic Music Generation using Recurrent Neural Networks (CS 224D: Deep Learning for NLP, Spring Quarter 2015) [code][video]

• Semi-Supervised Learning for Sentiment Analysis (CS 229: Machine Learning, Fall Quarter 2014)

• On the Riemann Hypothesis and Hilbert’s Tenth Problem (MATH 391: Logic Research Seminar, Winter Quarter 2012)