Joshua Levy

LinkedInGitHubThesis lab


HOMETOWN
Walnut Creek, California, USA

LANGUAGES
English

THESIS MENTOR

FAVORITE ALGORITHM
Graph Convolutional Neural Network & Poincare Embeddings

UNDERGRADUATE INSTITUTION
University of California Berkeley
Bachelor of Arts in Physics

PREVIOUS WORK EXPERIENCE
Zymergen
Software Engineer Intern
Joint Genome Institute
Software Developer
San Francisco Department of Health
Public Service Aide

DARTMOUTH QBS PROFILE

THESIS TITLE
Cross-Media Integration, Deep Learning and Highly Interpretable Models

CAREER TRACK(S)
Academica, Industry

TECHNICAL EXPERTISE
Python, R, SQL, Bash/Shell, High Performance Computing/Supercomputer Clusters, C++, MATLAB, Javascript, Deep Learning and Compter Vision, Genomic Analyses, Plotly, Keras, Docker

EXTRACURRICULAR ACTIVITIES
Dartmouth Graduate Student Council QBS Representative 

 


QBS PhD Experience

Electives
QBS 108: Applied Machine Learning

Publications

Poster Presentations

Oral Presentations


My Interests

"I'm interested in developing cutting-edge deep learning models that make strong biomedical predictions for molecular, histopathological, textual, and imagery data. Also, I'm looking to generate simple explanations for these complex predictions, so that members of the medical community will be quick to trust and adopt these models."

RESEARCHER I ADMIRE
Paul Dirac
(The theory about the Dirac Sea tying quantum mechanics, special relativity, magnetism, and anti-matter together is mind-bending)

FAVORITE RESEARCH PAPER
Neural Ordinal Differential EquationsModel Predictions
(I couldn't decide, so here are my two favorite NIPS papers)

Why QBS?

"I appreciate that QBS is highly collaborative and multidisciplinary program. Something that is really important to me is being able to combine ideas from multiple fields to solve a complex problem. I think the mental heuristics acquired from studying one domain, such as my original field of study, Physics, can be used to quickly solve complex problems in another domain, such as understanding the complexities of high dimensional data. The biomedical field is filled with many of these problems, and in need of someone who is not afraid to incorporate ideas or work with others from multiple fields of study to solve intractable problems."

FAVORITE QBS CLASS
QBS 100: Molecular Basis of Human Health & Disease

Life @ Dartmouth

WHEN I AM NOT IN LAB
"I'm outdoors, soaking in the sunlight (when it's not Winter), hiking with friends, working out at the gym, starring and contributing to the open-source GitHub community or enjoying some gelato after a hard day's work."