Nathaniel Todd Email & Phone Number

Pittsburgh, Pennsylvania, United States

Pittsburgh, Pennsylvania, United States

• Successfully led effort to transition submersible system software from research repository to production• Effort included creating .deb software release for installation on x86 and embedded ARM linux systems• Designed and created web-based dashboard for non-operational needs such as gathering diagnostic info, configuring different system settings, and downloading/uploading files from the system• Created automated CI/CD pipeline for web-based dashboard for portability and seamless re-use among other company projects• As part of deployment, worked to cross compile code for ARM based systems and package install into Debian file

Pittsburgh, Pennsylvania, United States

• Continued support of M2NS with a focus on reliability in operational conditions for final demonstration• Designed perception features leveraging point-cloud techniques such as ICP, Model Sample Consensus, Surface Normal Estimation, etc• Utilized point-cloud techniques both in Point Cloud Library (PCL) and OpenCV• Integrated Docker into development workflow for improved portability and repeatability• Optimized DDS networking to balance reliability on complex low bandwidth network hardwareDuring the acquisition of RE2 Robotics by Sarcos Technology and Robotics Corporation, I continued my support of the M2NS program and key role in design iterations centered on reliability in challenging operational conditions for the final demonstration. Notably, I conducted prototyping for perception pipeline updates, with both OpenCV and PCL, leveraging point cloud techniques/algorithms such as ICP, Model Sample Consensus, Surface Normal Estimation, and others which significantly improved the system's performance in dynamic environments. Furthermore, I integrated Docker into our development workflow forheightened portability and reliability across various environments and architectures.

Pittsburgh, Pennsylvania, United States

• Consulted on perception prototyping for RE2 aviation project for robotic baggage handling• Utilized stereo cameras to gather data and design object localization algorithms• Leveraged classical computer vision techniques including hough transforms, clustering, SIFT/ORB descriptors• Served as software lead of submersible system as part of the Maritime Mine Neutralization System (M2NS) program; a continuation of the DM2S program• Led software modularization for integration on embedded subsea nvidia jetson processors• Spearheaded perception code iteration for increased modularity and maintainability • Integrated submersible stereo camera• Iterated on perception algorithms incorporating stereo depth data, hough transforms, and more reliability to environmental conditions• Developed custom state estimation, filtering, and outlier rejectionDuring my promotion at RE2, I took on increasing responsibilities, consulting on the perception design on the aviation projects and being entrusted software lead responsibilities for the submersible system within the Maritime Mine Neutralization System (M2NS) program, a direct continuation of the DM2S initiative. As part of the aviation project, I prototyped localization/state-estimation algorithms with Python and OpenCV employing classical computer vision techniques such as hough transforms, clustering, and SIFT/ORB descriptors. Additionally, I led the software modularization effort on M2NS, enabling the use of embedded subsea NVIDIA Jetson processors. Furthermore, I led a large perception package update for better modularity and maintainability. Lastly, I integrated a submersible stereo camera and iteratively refined perception algorithms, incorporating stereo depth data, hough transforms, custom state estimation/filtering, and enhancements for increased reliability in challenging environmental conditions.

Pittsburgh, Pennsylvania, United States

• Prototyped perception algorithms using classical computer vision techniques - RANSaC, Image Guassians, Noise Filtering, Kalman Filtering • Performed Camera Intrinsic and Extrinsic Calibrations • Prototyped Algorithm Design in OpenCV using Python • Led final software design of perception package in C++ ROS1 using OpenCV • Performed Intrinsic and Extrinsic Camera Calibration, experimented with various solvers to address challenges of underwater camera calibration • Integrated autonomy system with 3rd part hardware and software • Optimized ROS communications for use on low bandwidth hardwareAs the Perception Lead on the Dexterous Maritime Manipulation System (DM2S) Project, I handled all computer vision design as part of the effort to integrate RE2 manipulators onto the Videoray Defender ROV while achieving high levels of autonomous capability. I prototyped algorithms to perform object detection and localization in 2D image data from the ROV camera and forward facing imaging sonar. Conducted Intrinsic and Extrinsic Camera Calibrations, ensuring precise and reliable performance.

• Developed fabric face detection utilizing image texture processing, principal component analysis (PCA), and support vector machines• Developed fabric orientation feedback algorithms to detect and correct orientation of fabric• Eliminated dropped frames in vision feedback with implementation of counter based image synchronization• Wrote automated tests to perform to verify camera sharpness, latency, and timing of machine vision system• Performed intrinsic and extrinsic camera calibrationDuring my tenure at Softwear Automation, I worked part-time as a computer vision software engineer while completing my Masters in Computer Science. I made significant contributions to the company's vision feedback system, enhancing its performance by eliminating dropped frames through the implementation of counter-based image synchronization. Additionally, I played a pivotal role in ensuring the reliability and accuracy of the machine vision system by writing automated tests to verify camera sharpness, latency, and timing. My expertise in computer vision techniques was put to use in the development of fabric face detection algorithms, employing image texture processing, principal component analysis (PCA), and support vector machines. Furthermore, I successfully designed and implemented fabric orientation feedback algorithms utilizing contour analysis and autocorrelation which resulted in an impressive 80% reduction in fabric orientation errors. Lastly, I conducted intrinsic and extrinsic camera calibration, further optimizing the vision system's precision and overall functionality. My experience at Softwear Automation was a valuable opportunity to apply my academic knowledge to real-world challenges in the field of computer vision and automation.•

Greater Atlanta Area

At GTRI's advanced concept CPTPO laboratory, I developed production java code. Our team worked to develop an android app for military groups to monitor mission equipment and issue commands. I developed features such as serial USB communication, UDP communication, and various new GUI elements.

Greater Pittsburgh Area

• Developed ROS nodes in Python and C++ • Used Tensorflow and OpenCV frameworks• Developed with Nvidia Xavier and Jetson TX2 platformsWorked with Carnegie Mellon researchers to bring agricultural research into a concept product. I worked on a robotic system that would gather agricultural data for various types of growers. During development, I applied GPU accelerated classical stereo vision, used various deep network architectures (Fast-RCNN, Faster-RCNN, and Single Shot Detection) for object detection, fine-tuned existing networks with proprietary training data, and integrated their vision systems with ROS on Nvidia TX2 and Xavier platforms.

Greater Atlanta Area

Worked as a Teaching Assistant for Georgia Tech's CS 4731/7632 Game AI Course. My responsibilities included grading Homeworks/Exams, holding office hours to assist students with class topics, and assisting the professor with proctoring or any other class duties.

Greater Pittsburgh Area

• Updated legacy VHDL and C code to support board redesigns• Used OrCAD and Visual Studio environments to make schematic, pcb, and software changes

Greater Pittsburgh Area

• Designed and prototyped legacy refresh of 10 layer PCB with OrCAD• Created project documentation to assist future interns with hardware development

Greater Pittsburgh Area

• Assisted in the design and configuration of drive control software • Collaborated with project engineers on field motor drive installation and commissioning• Compiled software summaries for New Product Introduction design reviews

Greensburg, Pa

"Mow'n'Go" was a small landscaping business I started and ran in high school. The business grew to 40 weekly customers and I led 2 regular employees. My summers spent with Mow'n'Go taught me valuable lessons about resource management, customer relations, and advertising.

Master Of Science - Ms, Computer Science, Gpa 3.9

Activities and Societies: Salsa Club - Vice President; KIDS Workshops - Instructor; Sailing Club - Member; CS Teaching Assistant; GTRI.

Bachelor'S Degree, Electrical And Electronics Engineering, Gpa 3.6

Activities and Societies: Pitt Robotics & Automation Society, Salsa Club, Pitt Interscholastic Ski Team Graduated Magna Cum Laude.