Inventory Management Core Data team (Aug 2022 - Feb 2024)
Helped design and implement rewrite of a large scale legacy application calculating supply chain data attributes for millions of Target items to improve its performance, scalability, and maintainability. Utilized an event driven microservice architecture as well as converted from a REST API to GraphQL for downstream consumers.
Fulfilled on call responsibilities/production support - fixed bugs and educated users, triaged and resolved several high priority incidents causing work stoppages at distribution centers.
Played key role in new team formation by creating both technical and non-technical infrastructure (e.g. build and deploy pipelines, application templates, documentation and working agreements, etc.) and took greater responsibility in product planning and mentoring interns/junior engineers.
Target Corporation - Software Engineer II
Supply Chain Business Data Authoring team (Nov 2021 - Aug 2022)
Built new back-end capabilities in a variety of data driven supply chain applications. All features included publishing to/consuming from Kafka topics, persisting data to databases, auditing, test coverage, and alerting/metrics. Collaborated with product owners to write stories and outline acceptance criteria, implemented design, managed releases on private cloud platform, and followed through with performance monitoring.
Examples include implementing business logic describing whether items are eligible for certain automation/fulfillment processes, building API allowing managers to set which stores can receive certain inventory pushes, and data ingestion app used to transform and publish data for trailer loading optimization.
In both legacy and modern applications - addressed tech debt (API migrations, major dependency updates, security vulnerabilities, etc.) and improved codebases (automated CI/CD steps, SSL cert renewals, secrets management, contributed to libraries, improved logs and observability, rewrote code, etc.)
CoolClimate Network - Software Developer
Jan 2021 - July 2021
Helped develop carbon footprint tool (coolclimate.org/calculator) used by both general public and businesses to achieve emissions goals. Improved service logic to increase API response accuracy, implemented UI features with React/Redux, managed deployments on AWS. Also worked as partial scrum master, organized sprints and communicated goals and progress to stakeholders
CAHL Research Lab - Software Developer & Machine Learning Research Assistant
May 2018 - Jan 2021
Worked on both research and development teams to build an AI driven academic planning platform for UC Berkeley, including a course recommendation feature. Implemented full stack features using Angular/Flask, helped train and productionize different machine learning models used by the system, refactored codebase to better follow OOP principles and modern architecture patterns.
Built automated data pipeline using Apache Airflow integrating campus APIs and registrar data dumps to retrain and version machine learning models and refresh MySQL data tables.
Published research paper demonstrating use of machine learning in live ed-tech platforms. Full paper PDF
Education
University of California, Berkeley
B.A. Data Science - High Distinction in General Scholarship (GPA: 3.93/4)
Awards and Publications
Conference Proceedings
Dong, M., Yu, R., Pardos, Z.A. (2019) Design and Deployment of a Better Course Search Tool: Inferring latent keywords from enrollment networks. In M. Scheffel & J. Broisin (Eds.) Proceedings of the 14th European Conference on Technology Enhanced Learning (ECTEL). Delft, The Netherlands. Springer. Pages 480-494. Full paper, Full paper PDF, Code
Undergraduate Research Fellow
Received $6,000 grant to conduct a self-directed research project under the guidance of a faculty mentor. Submitted work for peer review and presented published work at a research conference.
Projects
AI powered course planning agent
Proof of concept AI agent built using Langchain & OpenAI LLM to help students with course enrollment planning. Code
Personal website
Static React site template modified for personal use and practice with web app development and deployment. Built CI/CD pipelines to run tests, publish build, and deploy to a custom domain. Code
CA County Waste Management
Forecasting simulation predicting CA county waste production based on historic data and other engineered features. Scraped data from government websites, trained and evaluated different prediction models, created a geospatial heat map with optimal model projections. GIS Map, Code