Develop scheduling software for the Deep Space Network (DSN). Created a tool to automatically generate DSN requests for missions, reducing the time to do so by over 90%. Integrated a new CSP solver into the scheduling workflow.Created a fully autonomous federated observation planning system for retasking commercial smallsats using alerts from integrated science model. Mostly in Python. A subset of this system is running operationally to monitor volcanic activity. It has tasked >80 Planet SkySat scenes,scheduled >6,700 ECOSTRESS scenes, and identified >600,000 Planet Dove scenes of volcanic eruptions.Developed path planning software with C++ and ROS for AUVs to locate hydrothermal vents in a simulated 20000km2 domain. One algorithm discovered the vent in 80% of tested problems.Prototyped algorithms for onboard cloud avoidance using a look-ahead sensor. On historical data, our algorithm observed 80% cloud-free pixels compared to a 30% baseline.Built an internal domain-specific language with Python for the MEXEC planning and execution system as a replacement to xml. This is being used to plan activities for flight missions such as CADRE and EELS.

Teaching Assistant for CMSC421: Artificial Intelligence. I met with students to answer questions and provide feedback on assignments. I also graded student homework, programming assignments, and exams, while helping to administer the course.