LocationsSan Jose, CA
I'm currently a student at MIT splitting her time between research and school. I started out working in bioinformatics and moved to developing statistical algorithms to process big sparse datasets. Now, I'm working in machine learning and have a particular interest in the theory behind it. Specifically, I'm looking to research in optimization theory, linear algebra, and statistical learning theory.
Proficient in Tensorflow, PyTorch, Keras, and other machine learning frameworks within Python. Experienced in building large and complex models with custom features.
Proficient in R and MATLAB, including the image analysis and machine learning toolkits. Knowledgeable about building packages in R from the ground up.
Strong knowledge of algorithm design and analysis as well as fundamental concepts such as complexity theory and information theory. Basic understanding of cryptography and game theory.
Knowledgeable about statistics, especially as it pertains to machine learning algorithms. Experienced in studying and optimizing from a theory standpoint.
FastNorm: Improving Numerical Stability of Neural Network Training with Efficient Normalization. 1st author. Women in Machine Learning Workshop.
Systematic Analysis of Sex-Linked Molecular Alterations and Therapies. Co-1st author.
Prediction Propagation for Domain Adaptation. 4th author. Workshop on Equivariance, Invariance, and Beyond.
EMDomics: a robust and powerful method for the identification of genes differentially e xpressed between heterogeneous classes. 4th author.
I'm working on several different projects, from generative modeling to reinforcement learning. My intent is to design algorithms that are more efficient not only due to engineering efforts but also due to mathematical properties.
Implemented and refined a novel domain adaptation algorithm for NLP that uses center-masked convolutions and phased training to improve transfer.
Derived a purely mathematical acceleration of weight normalization that is more numerically stable and requires fewer computations. The runtime reduced from quadratic to linear.
Defined a notion of irreducibility in monoidal categories and investigate categories satisfying this criterion. Explored geometric interpretations of congruences in categories.
Developed a new statistical algorithm to detect differences in genetic expression between classes of patients. New method involves comparing the distribution of expressions instead of just the mean values (as was state of the art).
Built a framework to analyze differences between male and female cancer patients based on genetic markers and predict which treatments may be more effective in one gender over the other.
Used machine learning to predict which breast cancer patients would respond to chemotherapy based on non-invasive scans. Novel application of image analysis techniques in conjunction with model architecture.
Pursuing a major in Mathematics with Computer Science and a minor in Philosophy.
As research editor, I oversee the peer review process and maintain the standards of the journal.
I worked with 11 other students in my class to design our class ring, and I built our website.
I'm on the board of the Undergraduate Society for Women in Math (USWIM) and we host events for female mathematicians within and outside of our community.
I train regularly in mixed martial arts (MMA), and I love to ride my bike around.