Louis - CS student @ Umich

Where I've worked

Software Engineering Intern @ Lyft
May 2020 - July 2020

Built, scheduled and rolled out timely rental renewal and charge email communications for renters through a microservice architecture, reducing the cost to renters by up to $500 a week from the consequence system
Identified and successfully spearheaded the fix of a major pay-service migration issue prior to hitting production, preserving integrity of payment card information emailed to users through 300K+ weekly receipts
Redesigned the driver verification system for rental renewals in Python to prevent Lyft drivers deactivated for safety and compliance issues from accessing Express Drive vehicles
Piloted post-COVID restart of escalation consequences for overdue rentals in collaboration with Operations

Software Engineering Intern @ CME Group
May 2019 - August 2019

Enabled fault tolerance capabilities for gateways to the Globex trading platform on local deployments in order to directly provide prod environment-like conditions for engineers and speed up team’s SDLC by up to 20%
Engineered a system of three daemons voting on primary and backup state of gateways through TCP connections in Java to maintain integrity of the platform in case of a gateway failure
Slashed regression tests runtime by developing program able to evenly split up tests across AWS EC2 instances

Software Developer @ Proquest Medical AI team
January 2019 - December 2019

Delivered a medical Named Entity Recognition web service for pharmaceutical companies and researchers to optimize FDA compliance search for ProQuest clients
Built a bidirectional LSTM-CRF model for PubMed document tagging using ELMo embedding and PyTorch achieving 87% recall and 83% precision on medical corpuses
Integrated the NER model with a Flask + React web app supporting ProQuest’s search engine for documents

Machine Learning Intern @ Monaco Informatique Service
June 2018 - July 2018

Built a financial receipt classifying tool to automate administrative tasks using Computer Vision (OpenCV)
Created reusable Python scripts to perform data preparation (text cleaning, categorical encoding …) for NLP, supervised and unsupervised Machine Learning tasks
Boosted accuracy on classification predictions by up to 20% using probabilistic analysis and thresholding

Brakes Engineer @ Michigan Hyperloop
September 2017 - September 2018

Worked in a team of diverse engineers to design and build a pod competing at SpaceX Hyperloop Competition
Utilized Python & Matlab to simulate pneumatics system evolution through braking phase
Implemented a fail-proof pneumatics system ensuring braking effectiveness for a pod traveling at 300mph

What I've built

Honda Mobility

Built a detector for dangerous close-by drivers to alert users at risk using a statistics-based Python algorithm during Honda Mobility Hacks.

Python R Data analysis Statistics

KTPuild

Developed an efficient user-friendly alternative to the Make tool, configurable in JSON and able to start compilations concurrently.

C++ Linux Compiler JSON

TempIt

Created a web platform for college students to find flexible job opportunities in the hourly-market during MHacks 11.

HTML CSS Flask Firebase

CountUp

Customizable iOS app to keep track of personal usage for variety of personal products (originally thought for contact lenses).

Swift CoreData module

Neurons

Used the Hodgkin-Huxley model to represent neurons with electric circuits and estimate the speed of neuronal transmission.

Python Numpy Matplotlib Jupyter

Piazza Classifier

Developed a classifier for piazza posts to tag their topic automatically. Used bayesian probabilistic approach.

C++ CSV data Probability