Hire Rohan Z. - MLOps Engineer

Developing the ML operations for kea’s autonomous voice ordering system for restaurants

During the summer of my Master's program, I had the opportunity to intern at Verdant Robotics as a Machine Learning intern. Through my internship, I significantly contributed to the Verdant data processing pipelines and modules for model training, comparison, and deployment. Some projects I worked on included expanding the labeling pipeline to handle multiple crop inputs, implementing an ESRGAN to superresolve data for use in newer models, developing a Python based module for file transfer and archiving across the different servers employed at Verdant, and spearheading the restructuring of the Verdant dataset creation process for model training.

As part of my work as a graduate student in the Master of Science in Robotics program at CMU, I am conducting research under Professor David Wettergreen on a neural network to semantically segment coral reefs into coral, sand, and algae. I preprocess and break down both hyperspectral and multispectral imagery to then be processed through the network, as well as manually tune and adjust network parameters in order to improve segmentation performance. After training, I then apply an ergodic planner to determine resource efficient sampling trajectories for in person expeditions in order to improve upon the model predictions. I presented the first year results of my research at WHISPERS 2022 and had the opportunity to chair a session at the conference as well.