● Collaborated with CTO of BLDUP to identify, train, and deploy BERT machine learning model with 88.45% accuracy to classify permit data using permit descriptions to be displayed on company website
● Deployed model as a production-quality microservice as a Flask web API endpoint in Docker container
● Cleaned and formatted permit datasets from Boston, Austin, and San Francisco containing over 3,000,000 data points using Jupyter Notebook
Selected into 2021 Hack.Diversity Fellowship, a career leadership and workforce development program integrating Boston’s innovation, corporate, civic, academic, and philanthropic pillars
● Collaborated with astrophysics professor, Kenneth Brecher, to assist his personal physics projects
● Created SolidWorks models for spinning and orbiting objects to be used in astrophysics class demonstrations, and lens holders used for dichotomy research
● Used 3-D printer to provide professor Brecher with needed pieces for projects