Title:
RadNeuro Research Engineer
KBR is looking for a Research Engineer to work with a team to develop energy efficient, radiation tolerant and fault tolerant Neuromorphic processor for aerospace applications, with an emphasis on Lunar, deep-space, and Martian deployments. This involves first analyzing leading aerospace Neural Network (NN) applications and prototyping NN applications to determine requirements (e.g., size of NN, arithmetic precision required, frame rate required) and then mapping to representative state of the art Neural net accelerators and Neuromorphic Processors.
The RadNeuro Research Engineer will, within the RadNeuro team, develop prototypes for the demonstration of embedded Neuromorphic applications in NASA-relevant areas, such as small UAS, rovers, or nanosats for the RadNeuro project. Modern neuromorphic processors provide high performance while using minimal power for tasks like vision-based rover navigation, entry-descent-and-landing, sensor fusion, or on-board decision making. Within this task, we aim to demonstrate selected capabilities using neuromorphic co-processors (e.g., Google TPU, Intel Loihi, or Akida brainchip) with a prototypical implementation on a small existing platform (TBD).
The task involves the integration of sensors and cameras with the neuromorphic system in an architecture typically involving small, embedded boards (e.g., UP-boards, ARM-based like Raspberry PI, Nvidia Jetson, or similar) with the NN-processor connected, potentially together with some custom FPGA configuration. A small and simple controller will be implemented to drive the demonstration system. Small amounts of embedded software (C, C++, python) will need to be implemented.
***Must be a U.S. Citizen or Permanent Resident***
Required Education, Experience, & Skills:
Master’s degree in electrical engineering, computer science or related field2 + years of relevant experienceA mix of hardware and software integration and testing, especially associated with radiation tolerance for advanced processorsDevelopment of embedded software and low-level sensor interfaces on small platformsProgramming skills in C (and/or C++) and python and experience with frameworks like ArduCopter/ArduRover or ROSWorking knowledge in Field Programmable Gate Arrays (FPGA)Some experience in customizing existing platforms (UAS, Rovers) like mounting and wiring sensors, cameras, and control computers is necessary to facilitate a quick construction and assembly of the prototype demonstrators
Preferred Experience & Skills:
Working knowledge in Deep Neural Networks (e.g., Keras or similar frameworks)Test EngineeringHardware and software testingField Programmable Gate Arrays (FPGA)Technology maturationSystems integration, roboticsUAV, rover, and/or Nanosat lab work
(20min delay)