Lidar / Perception Team

Worked with Intel's Autonomous Driving Group (ADG) to develop algorithms for automated driving policy. Reinforcement learning methods and their application to this problem was our primary focus. We worked on applications for both the behavioral and motion planning stages of the system where the vehicle must decide what local goal is required and then how best to accomplish that directive. This work allowed us to benchmark the performance of state-of-the-art algorithms in order to directly communicate the strengths of Intel hardware to our customers.

Also worked on benchmarking and optimization of deep neural networks for object detection and localization for deployment on Intel vision accelerator hardware. Through this work, we also produced and submitted a patent application regarding methods for distributed object detection processing.

Additionally played a part in the analysis and optimization of a customer algorithm for occupancy grid mapping based on particle filtering. As a result of the success of our work, I received a division recognition award (DRA) for my contributions.

[Experience with TensorFlow, Keras, ROS, Python, C++]

Developed software for an airborne surveillance tracking system that enabled the detection of areas of shadow and occlusion within 2D imagery. My algorithm utilized LIDAR data (automatically selected from a large set based on the geographical location of the photo) to generate a 3D rendering of the visible scene using C++, OpenGL, and OpenCV, in which we could cast virtual shadows then refer the results back to the original image. This environmental information then represented one part of a larger "context-aware" multi-target tracker system being developed at the company for DoD airborne surveillance applications.

Ran experiments to optimize the time and material usage of different processes in the chemical vapor deposition (CVD) operations on silicon wafers. This work additionally contributed to improving the availability of the tools. I also worked with my team to characterize newly installed tools and to establish guidelines for implementing new procedures.

Worked as an undergraduate research assistant on a research project under the supervision of Dr. Rich Radke. The product was a computer vision system with the ability to track the movement of objects traveling through the field of view of multiple (distorted) camera lenses. Video streams from each camera were processed separately, then accordingly stitched together into a single mosaic video showing the entire scene, onto which bounding boxes were superimposed around subjects of interest.