I am a fourth year PhD student in the UCI department of Statistics supervised by Professor Pierre Baldi. My research focusses on hyperparameter search for machine learning models. I am also actively developing SHERPA, a hyperparameter search Python library. I am applying this work predominantly in the field of Neutrino Physics. I received my Bachelor of Science in Physics from Queen Mary, University of London and a Master of Science in Statistics from Imperial College London. I have done internships with Honda Research Institute and Apple. At Apple I worked on building a hyperparameter search system under the Platform Architecture Team. Resume.


Approximate Inference for Deep Latent Gaussian Mixtures
Convolutional Networks for Neutrino Energy Reconstruction


Sherpa is a black-box optimization system developed in the Baldi-Group at UC Irvine specifically for tuning hyperparameters in deep neural networks. Its goal is to assist the neural network researcher in finding a good set of hyperparameters. Sherpa uses a dashboard to visualize trials to the user and interfaces with job scheduling systems such as Sun Grid Engine or Slurm. A variety of algorithms are currently implemented ranging from grid search to Bayesian optimization, and Population Based Training (Jaderberg et. al 2017). Sherpa can be found here.

Paper Summaries

These are notes on my favorite papers.

Understanding Deep Learning Requires Rethinking Generalization
Deep Rl Through Policy Optimization (tutorial)
Hyperband: Bandit Based Configuration Evaluation For Hyperparameter Optimization
Label Free Supervision Of Neural Networks With Physics And Domain Knowledge


I am currently the teaching assistant for STATS 7 (Introductory Statistics). Other classes that I have TA'ed before are STATS 240 (Multivariate Analysis - Master level) with Dr. Zhaoxia Yu and STATS 111/202 (Categorical Data Analysis - Master level) with Dr. Stacey Hancock. See below for materials.