Jan 2023 – May 2023 · 4 mos · Raleigh, North Carolina, United States · Hybrid

• Developed inherently interpretable machine learning through instance-based learning.

May 2022 – Aug 2022 · 3 mos · Durham, North Carolina, United States

• Developed and validated ML-driven models for plant phenotypic scoring, saving the company more than $250,000.

Jun 2021 – Dec 2025 · 4 yrs 6 mos · Raleigh-Durham-Chapel Hill Area

• Built an automatic phenotyping platform for cotton researchers using object detection to detect narrow or wide cotton fiber tips, saving significant amounts of manual labeling time. Leveraged Google Colab, Roboflow, Detectron2, and a Faster R-CNN backbone architecture to create the application.
• Developed an AI-assisted labeling pipeline for cotton fiber tips using Detectron2 and Roboflow, reducing the labeling time threefold compared to fully manual labeling.
• Published results of cotton fiber tip analysis in the Applications in Plant Sciences journal.
• Train custom CNNs using TensorFlow and Keras to predict average cotton fiber tip diameter within ~0.5 microns of human measurements.
• Experiment with creative data augmentation techniques to handle imbalanced cotton diameter data.
• Build an application for blueberry breeders that utilizes object detection to automatically and precisely measure blueberry size, where the differences between the predicted mean and human-measured mean follows approximately a normal distribution with mean of 0 and standard deviation of 1.
• Coordinate an interdisciplinary team to label blueberries in Roboflow.
• Explore the use of NeRF techniques for data augmentation in order to improve real-world plant detection.
• Collaborate with robotics researchers to build custom imaging hardware for 3D modeling of plants.

Mar 2019 – Jun 2020 · 1 yr 3 mos

• Developed semantic segmentation algorithms using Mask R-CNN to measure dogs' tail positions with greater accuracy and less effort than state-of-the-art veterinary behavior techniques.
• Published results of dog tail analysis at the International Conference on Animal-Computer Interaction.

Jun 2017 – Aug 2017 · 2 mos

• Wrote op-eds with topics ranging from the Asian-American experience to why every student should learn how to code.
• Wrote columns for the news, sports, and features sections.
• Collaborated with the Technician editorial staff to find story ideas.
• Conducted interviews for various articles.

May 2017 – Aug 2017 · 3 mos · Cary, NC

• Tested the beta version of SAS Visual Analytics 8.1 by analyzing software performance.
• Created data visualizations and reports on sustainability data at SAS.