HMC Clinic Program: Rockerbox

Project Manager
September 2021 - Present
As part of HMC’s Clinic Program this academic year, I have served as the project manager of the team working with Rockerbox, a marketing attrition company. Our team is building an unsupervised anomaly detection model on time-series marketing data through a dynamic regression model and a long short-term memory (LSTM) autoencoder.


Software Engineer Intern
May 2021 – August 2021
I interned at Etsy on their search query understanding (SQU) team. At the time, the SQU team faced the issue of inappropriate queries being suggested to users and trending as top queries. I developed and implemented the solution to this problem, including constructing the training dataset with SQL. Using Tensorflow, I developed a DistilBERT language model. In addition to supplying the text of the query, I incorporated numerical query metrics into the model’s inputs. Using the Google Cloud Platform for training and storage, the model achieved 91% accuracy, and the SQU team will be moving the model to production. Additionally, I retrained an existing transformer-based model that classifies search queries as broad or direct, where I increased accuracy by 9% and reduced model volatility.


Software Engineer Intern
May 2019 - August 2019
I interned at Viasat on the NextGen Tactical Data Links team. Our intern team of three designed and built a heads-up display that connected to a Link 16 radio network in order to receive, display, and send information. I built a REST API using CherryPy Python Library to enable communication between the display and the Link 16 radio network. I implemented a global runtime manager in C# to handle distribution of data into assets, allowing for live updating of heads-up display. The project was presented by Viasat at the Association of the United States Army (AUSA 2019) conference.