Career Timeline

This was a project during my volunteering with the Marine Megafauna Foundation (MMF). In collaboration with Dan, we prototyped and built a website to help dive centers interested in contributing to citizen science highlight their data.

The data is visualized using Plotly and Dash. Three Plotly cross graphs, meaning what you select in one graph is reflected in the other two, make it possible to explore the data. This fledgling project is currently tracking 14 species of megafauna in the Inhambane province of Mozambique, and we are working on expanding it to other locations.

I modeled the behavior of aircraft in the airport pattern as a hidden Markov Model (HMM) whose parameters are learned from real-world radar observations. Then I determined optimal advisories to reduce the risk of collision by formulating the problem as a partially observable semi-Markov decision process (POSMDP), and solved it using Reinforcement Learning techniques.

In order to address the computational complexity of solving the problem, I used different approximation methods including exponential sojourn times, phase-type distributions, online algorithms, and particle filters for belief estimation.

Using Beautiful Soup, I scraped weather data relevant to aeronautical operations (winds, temperatures, clouds, rainfall, etc.) for different locations.
I computed cumulative histrograms, and linked the different criteria to each other. Then using an interactive display (built on Bokeh), I made it possible to adjust criteria to determine the number of flyable days for each location, and visualize the sensitivity to the different parameters.

I explored the use of an on-board camera to provide altitude estimation as a helping tool that could be used for training new pilots. I collected data from a cockpit view of an aircraft, and implemented a pipeline which detects runway edges and estimates the aircraft altitude.
Final results were promising but noisy (could benefit from Kalman filtering with some additional sensors)

At Zee.Aero (now cora.aero), I built a pilot simulator to experiment with control laws and inceptors, as well as for flight test pilots to train ahead of real life flights. This involved connecting hardware-in-the-loop avionics with a 6DOF aerodynamics simulation, and displaying the aircraft's behavior on a curved screen. This required calibrating the projectors, minimizing and quantifying the lag between inceptors and display, building tools to emulate failures, etc.

As an Aerospace Engineer, I participated in the flight testing of a proof of concept eVTOL aircraft. This included numerous flight-test campaigns at NASA facilities (Armstrong Flight Research Center (formerly Dryden) and Ames Research Center) in which I helped plan, execute, and monitor the performance of the aircraft and its subsystems.

As one of the early employees of Kitty Hawk (formerly Zee.Aero), I participated in the design, construction, aerodynamic and stability analysis, control synthesis, and flight testing of the proof of concept for eVTOL aircraft. This included working on subscale models, both in free flight but also collecting windtunnel data.

Using high power LEDs and computer vision, we determined the relative position between two UAVs in formation flight. The goal of the project was to demonstrate power savings when operating in the wake of an aircraft.

The Swift UAV is a foot-launched tailless sailplane that was converted to an electric UAV. This project was going to be my original PhD Thesis: building a sub-scale version of the aircraft in order to understand effects of scaling on the aerodynamic performance and dynamic response of sub-scale models.
I generated the aerodynamic stability derivatives used in the 6DOF simulation, and participated in the design of the avionics architecture for the Swift UAV.

As part of a group of Stanford Aero/Astro students, we designed, built, and tested small UAVs that we used to set the world altitude record for ’autonomous electrical UAV under 5kg’. We launched the UAVs at NASA's Dryden (now Armstrong) flight research center.

I created a remote assembly scenario-concept using a Haptic pen to control an industrial Kuka robot, a .NET wrapper for C++ APIs, and 3D stereo visualization for the scene.