Resume
Education
Harvard University | Cambridge, MA | Dec 2019 | M.S. in Data Science | GPA: 3.96/4.0 |
Related Coursework: Data Science 1+2, Generalized Linear Models, Artificial Intelligence, Parallel Computing
Brown University | Providence, RI | May 2018 | B.S. in Astrophysics w/ Honors | GPA: 3.81/4.0 |
Related Coursework: Computational Physics, Computer Systems, Matlab Programming, Statistical Inference, Quantum Mechanics, Differential Geometry, Cosmology, Partial Differential Equations, Linear Algebra
Work Experience
Summer Data Scientist | Boston Consulting Group | Gramm Division | Summer 2019
- Aided in construction of feature selection and attribution models to identify drivers of client’s spending
- Designed analytical framework to interpret selected features in context and share insights with client
- Built anomaly detection pipeline, capitalizing on forecasting and statistical process control methods
Graduate Teaching Fellow | CS 205: Parallel Computing | Harvard University | Spring 2019
- Led lab sections, held office hours, and graded homework for graduate course in parallel computing
- Taught students how to use multiprocessing, MPI, GPU, Spark, etc. on cloud computing infrastructure
- Received positive feedback from students about lab sections and homework help, earning a TF rating of 4.8/5
Applied Computation Internship | ISCR Program | Lawrence Livermore National Laboratory | Summer 2018
- Trained random forest machine learning models to predict and avoid simulation failure from geometric data
- Identified redundancy in the original feature set and analyzed behavior of reduced feature models
- Explored alternate labeling schemes to improve model performance and generalizability
Design Physics Internship | HEDP Program | Lawrence Livermore National Laboratory | Summer 2017
- Designed ICF nuclear fusion simulations to test NLTE physics integration into multi-physics code
- Wrote Python visualizations of spectral data to analyze NLTE effects on radiation energy density and opacity
- Exposure to HPC, deep learning, ALE mesh algorithms, parallelization, and code efficiency methodologies
Data Science & Coding Projects
- Project goal: build a deep learning model to classify ultrasound videos as one of five pathologies
- Developed, tuned and tested the efficacy of CNN+RNN combo, 3D CNN, and transfer learning
- 3D CNN yielded promising result: 75% accuracy on hard-to-identify Transposition of Great Arteries (TGA)
- Project goal: take a base image template and create a model to automatically generate a funny caption
- Scraped memegenerator.net for 805 base image + 70,000 unique captions, and filtered for harmful language
- Built an intricate encoder-decode architecture to successfully generate humorous results from image features
Spotify Playlist Recommendation | Data Science 1 | Fall 2018
- Project goal: use Spotify playlist data to create recommendation system that suggests relevant tracks
- Implemented and compared kNN collaborative filtering, k-means clustering, and ALS matrix factorization
- Achieved 23% R-precision on hold out set w/ ALS factorization compared to baseline of 7% w/ kNN method
- Project goal: develop RL agent that identifies best time to purchase stock within a given time frame
- Scraped daily stock price history from Yahoo! Finance and engineered an original Markov Decision Process
- Employed and analyzed Q-learning, as well as function approximation w/ linear weights + neural networks
Awards
Software
- Python | Matlab | LaTeX | Git | Bash
- C | SQL | R | Spark | Bash| AWS | Tableau