CV

B.A. in Computer Science + Cognitive Science

Columbia University

• Computer Science + Cognitive Science
• Coursework --> Data Structures in Java, Advanced Programming in C, Linear Algebra, Intro to Databases, Parallel Functional Programming, CS Theory, Intro to Cog Sci, Cog Neuroscience, Visual Neuroscience, Computational Genomics.

Software Developer, Researcher | Co-Founder

Ocular Diagnostics, Inc.

• Developed ML model with 86% AUROC to predict risk of ADHD using pupil-size variation over time-series—used Scikit-learn, NumPy.
• Implemented nested cross-validation, univariate feature selection, Fourier transforms, spectral analysis, velocity/acceleration analysis, SVMs.
• Engineered 783 pupillometric features; used nonparametric statistical tests + feature visualization. Developed Python package pupil-metrics.
• Developed C++ desktop GUI + low-cost VR headset that streams infrared USB cameras to AWS server for analysis + use in clinical trials.
• Founded startup backed by Soma Capital ($100K uncapped investment), $30K in grants. Published research in Nature Scientific Reports.

Software Engineer Intern

Helix (leading population genomics Series C startup)

• Building health chatbots.

Applied AI/ML Software Engineer Intern

Nokia Bell Labs

• Implemented post-processing logic of end-to-end ML workflow for dynamically sending data to IoT environmental sensors-used Python.
• Developed rule-engine controller to define post-processing rules + group controller to send data to sensors in groups—tested with Docker.

ML/Software Engineer Intern

Symbotic

• Developed partial K-Means clustering based method of surface normals for detection of planes from 3D pointclouds-used OpenCV, Open3D.
• Analyzed prior RANSAC-based algorithms for plane detection; parsed and extracted methods from prior developed code.

Computer Vision + Deep Learning Intern

Icahn School of Medicine at Mount Sinai

• Finetuned ResNet network with mammographic images to predict breast cancer risk within 1, 1+, 3+ years with ~60% AUC--used PyTorch.
• Created saliency maps based on model outputs to model network intuition. Established proof-of-concept and basis for future research funding.

Computer Vision + Deep Learning Intern

New York University - Video Lab

• Developed neural networks for auto-segmentation of body and brain ventricle (BV) from 3D mice ultrasound images—used PyTorch.
• Trained VNet CNN to create coarse segmentation map of body and BV; created additional VNet to refine coarse maps with 93% Dice metrics.
• Helped optimize inference time of prior state-of-the-art sliding window based pipelines by 1000x. Published method in IEEE as co-author.

pupil-metrics

• A small package for computing pupil-size metrics over a time-series.

TrueDetect

• A Deep Learning-Based Fusion Framework for Detecting Fake News (Columbia Data Science Hackathon Fall ‘21)

GRNPar

• Parallel Inference of Gene Regulatory Networks Using Boolean Networks and Information Theoretic Approach

• 1st Place Award @ 2022 Columbia Undergrad Venture Competition ($15K grant)
• 2nd Place Award @ 2022 Columbia Venture's Technology Challenge ($15K grant)

• Milton Fisher Scholarship for Innovation and Creativity

• Published + Speaker @ 11th ACM International Conference on Bioinformatics, Computational Biology, and Health Informatics
• Published + Speaker @ IEEE/ITU International Conference on AI for Good