Machine Learning for State Estimation
at ORAU Workforce Solutions

Date Posted: 3/23/2019

Opportunity Description

Title of Research Opportunity: Machine Learning for State Estimation and Decision Propagation ARL requires one full time engineering research scientist/engineer for a post-doctoral fellowship to support a program in Network of Networks to enable novel methods for agent learning, adaptation, and model distribution in highly heterogeneous environments where humans are coupled to machine decision agents. Specifically, the opportunity will develop novel theories, develop experiments to validate, and hardware to implement solutions to the problem of: Given the advances made by the Google DeepMind project (and others) how can we enable both hierarchical and deeply integrated Human-in-the-Loop (HIL) reinforcement, transfer learning for heterogeneous agents, and extend these methods from simulation demonstrations to hardware in the loop mixed systems? This person will be expected to lead their own research efforts but participate within a highly collaborative research group. This person will be expected to publish first author efforts in peer reviewed literature; contribute technically to peer reviewed literature in diverse areas within and outside of the team; and, develop experimental and transition efforts across the team. This opportunity will involve a mix of skill sets ranging including a deep understanding of various machine learning, transfer and reinforcement learning techniques, developing theories supporting multi-agent learning, and transfer of learned behavior across heterogeneous agents. This project is expected to research fluidly in python, Linux, ROS, Matlab, C/C++. Required Skills and Knowledge: • 5+ years of experience in Python and C/C++ or other scientific languages specifically geared towards Machine Learning; familiarity and experience with prototyping in MATLAB • A deep understanding of many different (machine, transfer, reinforcement, DCNN) learning techniques, and the demonstrated ability to bring state of the art methods from the literature in house. • Fluency in Linux and Windows development environments • Rigorous understanding of and experience with generating and executing experimental designs using large data sets. Experience implementing (deep) learning algorithms on multi-modal datasets • Good communication and presentation skills, and a demonstrated track record of peer-reviewed publication. • Ability to generate technical documentation of results and software; proficiency with integrated development environments, revision management systems such as Git/svn, and other tools for building large applications Desired Skills and Knowledge: Recent graduate with a PhD in control systems, electrical, mechanical engineering, computer science, material science, mathematics, physics or other appropriate discipline. Description: ARL requires one full time engineering research scientist/engineer for a postdoctoral fellowship to support a program in Network of Networks to enable novel methods for agent learning, adaptation, and model distribution in highly heterogeneous environments where humans are coupled to machine decision agents. Specifically, the opportunity will develop novel theories, develop experiments to validate, and hardware to implement solutions to the problem of: Given the advances made by the Google DeepMind project (and others) how can we enable both hierarchical and deeply integrated Human-in-the-Loop (HIL) reinforcement, transfer learning for heterogeneous agents, and extend these methods from simulation demonstrations to hardware in the loop mixed systems? This person will be expected to lead their own research efforts but participate within a highly collaborative research group. This person will be expected to publish first author efforts in peer reviewed literature; contribute technically to peer reviewed literature in diverse areas within and outside of the team; and, develop experimental and transition efforts across the team. This opportunity will involve a mix of skill sets ranging including a deep understanding of various machine learning, transfer and reinforcement learning techniques, developing theories supporting multi-agent learning, and transfer of learned behavior across heterogeneous agents. This project is expected to do research fluidly in python, Linux, ROS, Matlab, C/C++. Required Skills and Knowledge: • 5+ years of experience in Python and C/C++ or other scientific languages specifically geared towards Machine Learning; familiarity and experience with prototyping in MATLAB • A deep understanding of many different (machine, transfer, reinforcement, DCNN) learning techniques, and the demonstrated ability to bring state of the art methods from the literature in house. • Fluency in Linux and Windows development environments • Rigorous understanding of and experience with generating and executing experimental designs using large data sets. Experience implementing (deep) learning algorithms on multi-modal datasets • Good communication and presentation skills, and a demonstrated track record of peer-reviewed publication. • Ability to generate technical documentation of results and software; proficiency with integrated development environments, revision management systems such as Git/svn, and other tools for building large applications Desired Skills and Knowledge: • Experience optimizing objective functions associated with machine learning algorithms • Experience with multi-agent modeling • Familiarity with Robot Operating System (ROS), Apache Thrift, Lightweight Communications and Marshalling (LCM), or other commonly used transport protocols for distributed and embedded computing environments • Experience with implementing and testing novel multi-sensor fusion and state estimation algorithms for robotics • Familiarity with Operator Theory. • Familiarity with IRB procedures and current CITI training. • Proficiency with integrated development environments, revision management systems such as Git/svn, and other tools for building large applications • Understanding of, and experience programing and integrating embedded systems. • Hands on experience with various robotic platforms and/or wearable sensor systems. • Experience with implementing and testing multi-sensor fusion and state estimation algorithms for robotics. • Practical understanding of experimental statistics including statistical experiment design, data analysis, validation and verification. Environment and Team Description: The Army Research Laboratory’s Electronics for Sense and Control seeks to enable revolutionary advances in three interconnected research spaces: 1) navigation and localization, 2) small scale autonomous systems, and 3) human physiological state monitoring. While an individual will have a specific project, it is expected that s/he will contribute across the full research space. The group’s approach includes the development of novel, distributed state estimation methodologies, linear and nonlinear controls integration, computer science and hardware engineering to realize a generalized framework for information acquisition and fusion in uncertain environments. The ESC team is comprised of a diverse technical team with highly interconnected projects. We foster an environment where hard work is rewarded; integrity is respected; and, insight, wherever it may come from, is valued. We honestly believe that a diversity of opinions should not only be respected, but it is expected, and we think that it is the only way to enable revolutionary advances. Team members each bring a unique skill set that we expect them to apply across completely novel applications. Individuals are expected to be highly self-reliant, and simultaneously capable and willing to collaborate across disciplines in dynamic team projects. The pace of projects is very fast; the expectations are exceptionally high; but, foremost, we value each team member, and advocate for a healthy balance in everyone’s life. We are looking to recruit new individual(s). They should be able to not only contribute from day 1, but be willing to learn from day 1. All eligible applicants including veterans, wounded warriors, and those with non-traditional educations are encouraged to apply.

Opportunity Snapshot

About Us

Grooming future leaders in science and technology requires enhancing the skills, knowledge and experience of workers early in their careers. To that end, ORAU (Oak Ridge Associated Universities) assists in connecting the best and most diverse students, recent graduates, faculty and professionals with world-class fellowships, internships and jobs, whether in national laboratories, research institutions, federal government offices or private sector R&D departments.

ORAU works with agencies such as the U.S. Army Research Laboratory (ARL) Research Associateship Program (RAP) allow Postdoctoral Fellows, Journeyman Fellows (undergraduate and graduates students and recent graduates), Senior Researchers, and Summer Faculty engage in research initiatives of their own choice, that are compatible with the interests of the government and will potentially contribute to the general effort of the ARL. We work with the Environmental Protection Agency (EPA) to place recent graduates in full-time and part-time jobs in the Office of Research and Development at EPA under the National Student Services Contract. ORAU also works with Center for Medicare and Medicaid Innovation.

Research opportunities include, but are not limited to the following disciplines: Aerospace Engineering, Anthropology, Archeology, Biology, Biochemistry, Biological Engineering, Biomechanical Engineering, Biomedical Engineering, Chemical Engineering, Chemistry, Computer Science, Computer Engineering, Data Science, Electrical Engineering, Environmental Health Risk Assessment, Environmental Science, Entomology, Epidemiology, Ergonomics, Geology, Health Education Mechanical Engineering, Materials Science, Mathematics, Nanotechnology, Photonics, Physics, Public Health Economics, Public Health Policy, Toxicology, and more.

If you are ready to take the next step in your career path, then we are ready to help you!

What is a Talent Network?

Joining our Talent Network will enhance your search and application process. Whether you choose to apply or just leave your information, we look forward to staying connected with you.


Why Join?

  • Receive alerts with new opportunities that match your interests
  • Share opportunities with family and friends through Social Media or email