Feb 2024 – Sep 2025 · 1 yr 7 mos

Aug 2022 – Dec 2022 · 4 mos · Remote

• · Enabled usage of satellite imagery in change-detection software by writing low-level software to parse raw
• sensor data from a satellite to create registered GeoTIFF images required by GIS software
• · Analyzed sensor characteristics and implemented several processing, filtering, and estimation algorithms to
• improve data quality, improving efficacy of change detection while reducing false positives
• · Collaborated across teams (scheduling and control and remote sensing) to produce imagery of several
• floods and hurricanes in real-time, which was given to the Federal Emergency Management Agency (FEMA)
• · Project website: https://www.ll.mit.edu/r-d/projects/disastersat

May 2021 – Aug 2021 · 3 mos · Boston, Massachusetts, United States

• Developed geospatial machine learning model to predict earthquakes using ionospheric data with 87.7% accuracy
• Identified, designed, and tested diverse architectures to determine an effective model for sparse and high-dimensional data
• Identified and extracted features of complex dynamical systems to learn efficiently from limited data
• Worked on a team to write readable, extensible, and well-documented code for future development in earthquake research, and gave a tutorial to onboard new team members onto the project
• Co-authored paper pending publication at AGU conference, Fall 2021: “Forecasting Earthquakes from Preseismic Ionospheric Signals Using Machine Learning – Refinements and Advances”

Aug 2021 – Jan 2022 · 5 mos · Atlanta, Georgia, United States

• Coordinated and supervised TA efforts, designed course content, and taught the principles of robotics software to a class of 150 students as a head TA of CS 3630 at Georgia Tech.

May 2020 – May 2023 · 3 yrs · Atlanta, Georgia, United States

• Designed anticipatory trajectory optimization software to enable natural, real-time human-robot collaboration
• Refactored existing lab code to improve readability and collaboration, enabling faster onboarding and extension for future research in human-computer interaction
• Co-authored IROS conference paper: “Anticipatory Human-Robot Collaboration via Multi-Objective Trajectory Optimization”
• Applied a multidisciplinary approach to solve incidental research problems, including contributing to a custom joint-space velocity controller for the JACO robotic arm
• Utilized state-of-the-art nonlinear solvers to reduce runtime of the optimization algorithm by 3 orders of magnitude, making it possible to improve trajectories in real time
• Implemented trajectory optimization pipeline on a physical robot and ran a human-robot collaboration study suggesting that users preferred my method over state-of-the-art baselines

Jan 2020 – Aug 2020 · 7 mos · Atlanta, Georgia, United States

• Designed terrain classification computer vision algorithms to enable path planning for a bipedal robot
• Initiated new effort to use unsupervised learning to learn new terrains online, including authoring and submitting a proposal for funding

May 2019 – Aug 2019 · 3 mos · Chantilly, Virginia, United States

• Investigated applications of convolutional neural networks to vision tasks for robotic refueling of satellites
• Employed state-of-the-art deep learning methods to retain high accuracy under challenging lighting conditions and clutter in a proof of concept by pre-training neural networks on synthetic data
• Authored two papers: one white paper on CNNs and one research paper on a novel ML algorithm

May 2018 – Aug 2018 · 3 mos · Atlanta, Georgia, United States

• Worked in a team to write a compliant control algorithm for robotic arms, enabling a variety of applications in research and manufacturing.
• Consulted on other research projects to deploy compliant control algorithm in a variety of settings
• Wrote backend code for planning software to enable broad application by non-roboticists

Jun 2017 – Aug 2017 · 2 mos

Dec 2016 – Jun 2017 · 6 mos · Washington D.C. Metro Area

Jun 2016 – Aug 2016 · 2 mos