John Lee
Fremont,CA
Overview
2
2
years of professional experience
Work History
Research Intern
CBCL - Stanford University
05.2022 - Current
- Under Dr, Alison Marsden - REU NSF CDS&E award #2104831
- Prototyped probabilistic surgical guidance pipeline identifying and parameterizing pulmonary stenosis regions
- Constructed 3D pulmonary cardiovascular models using Simvascular, tuned boundary conditions with Nelder-Mead optimization, and solved 3D computational fluid dynamic and 0D lumped parameter network simulations using HPC clusters and multiprocessing.
- Introduced a novel linear correction method that efficiently optimizes 0D models to 3D solutions, solving problems with modeling important pressure drops in vessel junctions and reducing mean pressure and flow errors to less than 5%.
- Trained neural network surrogates simulating hemodynamic pressure and flow results within 1/20th of a standard catheter's margin of error.
- Implemented OOP interface code to work with Simvascular 0D lumped parameter network and solver components
- A paper is in works with intentions to file a patent.
Research Intern
Dept. Of ACMS - University Of Notre Dame
06.2021 - 05.2022
- Under Dr. Daniele Schiavazzi - REU NSF CAREER award #1942662
- Developed a multi-fidelity U-Net model for retinal vessel segmentation, quantifying the effects of low-fidelity image data on prediction capability in 12 varying fidelity compositions
- Identified novel rotation UQ method allowing for differentiation of U-Net models based on data fidelity
- Poster presentation at University of Notre Dame, COS-JAM.
- Attended meetings and presentations to learn project details
Education
Bachelor of Science - Computer Science / Applied Math And Statistics
University of Notre Dame
Notre Dame, IN01.2024
Skills
- Python - 55 years, C - 3 years, Knowledge in JavaScript, HTML, C, Java, Clojure
- PyTorch Ecosystem, Neural Networks, ML/DL
- Probability, Stochastics, and Optimization
- Distributed Systems and High Performance Computing
Projects
StockNet - A Distributed Stock Game
- Designed a distributed system that simulated a simplified stock market, involving multi-entity communication via TCP connections.
- Ensured crash robustness, analyzed tradeoffs in latency, throughput, and persistence, and quantified performance as we scaled up.
2048RL
- Implemented a 1-depth Monte Carlo tree search Deep Q Network, successfully using reinforcement learning to play the 2048 game.
BMinor Compiler
- Built components of a compiler (scanner, parser, type-checker and code generator) for the BMinor language using flex, bison, C and x86.
GenderPredictor
- Utilized CelebA Dataset and transfer learning from a pretrained face detection model (VGGFace2, Casia) to predict the gender of an individual.
- Used ethnically diverse dataset to quantify model biases resulting from training using ethnically white dataset.
Timeline
Research Intern
CBCL - Stanford University
05.2022 - Current
Research Intern
Dept. Of ACMS - University Of Notre Dame
06.2021 - 05.2022
Bachelor of Science - Computer Science / Applied Math And Statistics
University of Notre Dame
Similar Profiles
Hector CamatonHector Camaton
Cardinal Quarter Fellow at Stanford UniversityCardinal Quarter Fellow at Stanford University
Sarah DiBoiseSarah DiBoise
Deputy General Counsel Health Care at Stanford UniversityDeputy General Counsel Health Care at Stanford University
Alyssia DouglasAlyssia Douglas
Patient Financial Counselor at Stanford UniversityPatient Financial Counselor at Stanford University
Carlos Quevedo GracianoCarlos Quevedo Graciano
Lead Installer at Pivot InteriorsLead Installer at Pivot Interiors
Divya KandimallaDivya Kandimalla
SQL Support Engineer at Tata Consultancy Services (TCS)SQL Support Engineer at Tata Consultancy Services (TCS)