· AlphaDogfight DARPA Challenge Member of 1 of 8 teams selected to compete in a virtual competition designed to demonstrate advanced AI algorithms that canperform simulated within-visual-range air combat maneuvering. Georgia Tech Research Institute wants to demonstrate the power ofleveraging optimal control theory and reinforcement learning for path planning/trajectory optimization.

Performing relevant research in the field of optimal control, autonomy and machine learning.· Solution to the Minimum-fuel Powered Descent Guidance Problem using Non-convex Optimization via Adaptive Stochastic Search Currently contributing to the development of a control framework that uses deep neural networks to learn stochastic optimal controlpolicies leveraging nonlinear optimal control theory, by formulating the minimum-fuel powered descent problem as a pair of forwardbackward stochastic differential equations. This work is an extension of the work by Exarchos et. al. in “Non-convex Optimization viaAdaptive Stochastic Search for End-to-end Learning and Control.” The purpose of the project is to demonstrate that this novelframework is capable of handling non-convex optimization problems subject to control and state constraints.· Risk-Sensitive Gradient-Based Adaptive Stochastic Search Contributed to efforts to extend the work by Zhou in “Gradient-Based Stochastic Search for Non-Differentiable Optimization” tominimize the Conditional Value-at-Risk of cost functions subject to nonlinear stochastic dynamics in belief space. The algorithm wasshown to effectively handle stochasiticity in the form of uncertain initial conditions, system parameter uncertainty, and stochasticity inthe control channels when tested in simulated inverted pendulum, cartpole and quadcopter systems.· Alphapilot AI Drone Innovation Challenge Supported the software implementation of an Stochastic Optimal Controller to be deployed on a quadrotor platform for AutonomousDrone Racing as part of one of the nine teams selected to participate in the 2019 AlphaPilot Innovation Challenge. Main tasksperformed included extensive controller testing and troubleshooting on simulation, cost map generation in C++/CUDA, cost functiondesign, development and integration of an Extended Kalman Filter for state estimation.

· Supported the development of a quadcopter testbed with the purpose of replicating the dynamics of fuel sloshing in launch vehicles to demonstrate the capabilities of an Augmenting Adaptive Control (AAC) algorithm.· Performed extensive testing to determine important parameters of the system (such as percent throttle to propeller RPM and RPM to thrust relationships) and supported the development of a nonlinear dynamical model in MATLAB/Simulink.· Designed and implemented a PD controller for attitude control.· Implemented a Linear Quadratic Regulator (LQR) for translational and altitude control.· Carried out trade studies to maximize gain and phase margins while minimizing response time of the overall system.· Co-authored the paper “Suppression of Nonlinear Rotary Slosh Dynamics using the SLS Adaptive Augmenting Control System Demonstration on a Quadcopter Testbed”

· Supported NEA Scout, a secondary payload in the first SLS Launch that will fly by a near-Earth asteroid using a solar sail as its in-space propulsion system. · Characterized disturbance torques for different sail shapes. · Performed mission design for future launch opportunities using MALTO. · Conducted trade studies to identify alternative targets for the mission. · Conducted trade studies to identify optimal flyby conditions and maximize the time budget to take a picture of the asteroid during the science slew portion of the mission. Generated a visualization of the results using Celestia. · Developed an interfacing tool using Python to facilitate comparison between different mission design software. · Supported IMU testing.

· Supported system modeling efforts for the hardware and the flight software of Orion AA2’s Crew Module that will communicate with and control the actual subsystem hardware. · Created notional architecture models for the avionics flight software and hardware using SysML and UML. · Demonstrated that model based techniques can be used to support the integrated Software and Hardware design for future projects by successfully importing requirements from the project’s repository, capturing software design and hardware/software integration and demonstrating traceability between requirements and software design in SysML. · Supported system modeling efforts for HESTIA’s Environmental Control and Life Support System (ECLSS) to carry out design trade analysis. · Worked closely with ECLS’ water and air system specialists to capture a Mars Life Support Notional Architecture and the behavior of the system using parametric diagrams and state machines in SysML.

Serve as a tutor for Montgomery College students. Mainly calculus I and II clarifying concepts like differentiation, optimization, integration, series and sequences, etc. Also aided other students with classical mechanics, electricity and magnetism, waves and other physics concepts.