I am currently a Master's student at Duke University. I receieved my Bachelor's in math and computer science at Duke, where I was fortunate enough to do research on interpretable multimodal deep learning with Cynthia Rudin and dynamic factor analysis with David Dunson. I am mainly interested in statistics and machine learning, particularly in informed decision making, theory-driven methodologies, and their impacts in healthcare and finance.

I use this site mainly as a personal index and a quick reference of the topics I've studied so far. Here is a graphically friendlier representation of my website.

I like to play volleyball, basketball, and cook for fun. Check out my menu.

Reading Notes Publications Hourly Problem Read Papers Projects

Math

Computer Science

Statistics

Physics

Chemistry

Coming Soon

Biology

Coming Soon

Economics

Political Science

Coming Soon

Psychology

Law

Coming Soon

Anthropology

Coming Soon

Sociology

Coming Soon

Philosophy

Epistemology

History

Religion

Coming Soon

Literature

Coming Soon

Visual Arts

Performing Arts

Coming Soon

Ember

A machine learning Python package with a backend autograd engine in C++, built only with standard libraries. The primary data structures are the ember.Tensor and ember.GradTensor objects. It supports construction and training of deep neural nets, classical ML algorithms, Monte Carlo sampling, and classical hypothesis testing/confidence intervals. An advantage of this package is that rather than only supporting batch vector operations and matrix multiplications, we can also perform general contractions of rank-(N,M) tensors.

Aqua

A visualization package for trajectories on the Bloch sphere. Not started yet.

Blade

A JavaScript library that allows you to build blockchains from scratch and use them for whatever you want, built only with standard libraries. You can create a lightweight and flexible blockchain following the Bitcoin protocol, with support for hierarchical wallets, ECC/RSA encryption, transaction handling, and asynchronous mining with adaptive difficulty. I've recently got into Monopoly, and so I implemented an application of it by coding the board game up where all the transactions are on this custom blockchain.

Indigo

An options trading window and simulator. Not started.

protopnet-pytorch

End-to-End Implementation of ProtoPNet in PyTorch. To do: ProtoTree, STProtoPnet, Deformable ProtoPnet

word2vec-pytorch

End-to-End Implementation of Word2Vec in PyTorch. To do: GloVe

rnn-pytorch

End-to-End Implementation of RNN in PyTorch. To do: LSTM, GRU

gpt2-pytorch

End-to-End Implementation of GPT2 in PyTorch. To do: GPT3

rbm-pytorch

TO DO: End-to-End Implementation of RBM in PyTorch.

vae-pytorch

End-to-End Implementation of VAE, CVAE in PyTorch.

gan-pytorch

End-to-End Implementation of GAN in PyTorch. To do : DCGAN, Wasserstein GAN, CycleGAN

normalizing-flows-pytorch

End-to-End Implementation of Normalizing Flows (NICE, RealNVP, Glow, cGlow) in PyTorch.

Publications

Hourly Math Problem

Hourly Coding Problem

Math

Computer Science

Statistics

Physics

Chemistry

Biology

Economics

Political Science

Psychology

Law

Anthropology

Sociology

Philosophy

History

Religion

Literature

Visual Arts

Performing Arts