- Fine-tuned a Mask R-CNN model and Vision Transformers to segment clasts in images taken by the Perseverance Rover’s ZCAM- Conduct extensive data analysis on segmented clasts, measuring key features such as area, long/short axes- Scheduled to present the research findings at the upcoming Lunar and Planetary Science Conference (LSCP)

- Fine-tuned InceptionV3 for frost detection on the Martian terrain using HiRISE imagery.- Model detect frost on Martian surface scans with 92% accuracy- Worked with Python, TensorFlow, Keras, and Pytorch to create and implement the models.- Collaborated with experts in geology, climate science, and machine learning.