The National Security Sciences Directorate at Oak Ridge National Laboratory leads scientific and technological breakthroughs to confront some of the nation’s most difficult security challenges. We develop interdisciplinary applications needed for the security of our nation today and target our vision on how these challenges may manifest themselves in a decade or more.
Our research and development focuses on cybersecurity and cyber physical resiliency, data analytics, geospatial science and technology, nuclear nonproliferation, and high-performance computing for sensitive national security missions. We also enhance ORNL contributions to national security challenges by working closely with leading researchers at the lab in areas such as nuclear and chemical sciences and engineering, applied materials, advanced manufacturing, biosecurity, transportation, and computing.
We are currently seeking qualified applicants for a Senior R&D Staff position in Geospatial Artificial Intelligence within our Geographic Data Science Section and GeoAI group. The position requires strong skills in computer science, optimization and machine learning for high performance computing including representation learning, hyperparameter optimization and distributed deep learning for geospatial applications.
The position affords the unique opportunity to work with a talented interdisciplinary team of R&D professionals, build new research directions; advancing GeoAI systems for computing on multi-platforms, advancing stochastic network architecture search methods, developing network compression algorithms to enable deep learning computing on the edge, cloud and high performance computing environments. The Geographic Data Science Section develops sensor technologies, analytical methods and models that collect, integrate, analyze and derive value from spatiotemporal data.
Within the GeoAI group, you will develop novel methods for machine learning application in large scale space-time analytics and multi-modal data fusion, take advantage of the world’s most robust high-performance computing and and cloud computing platforms, aim for big geospatial science opportunities to support critical national security missions.
Major Duties and Responsibilities:
- Initiate, lead and perform independent and impactful R&D on an ongoing basis as evidenced by, for example, innovative S&T artifacts delivered to sponsors, publications, successful proposals, technical presentations and technology demonstrations, professional community engagement, inventions/patents/copyrights, etc.
- Publish in top tier conferences (including SuperComputing, ACM SIGSPATIAL)
- Develop and implement plans to exploit new opportunities and/or expand existing research efforts.
- Provide relevant input to assist the R&D Group Lead and Section Head with staff development while providing direct and impactful guidance to staff on regular basis.
- Lead by example through exercising scientific integrity in proposing, performing, and communicating research.
- Lead by example to ensure all work is carried out safely, securely, and in compliance with ORNL policies, standards, and procedures.
- Exemplify a commitment to excellence in research, operations, and community engagement, and work cooperatively to leverage scientific capabilities across ORNL.
- Advance the reputation of the group members and standing of the group as a whole through establishing collaborations, professional society leadership and involvement, and organization of technical events at major conferences.
- Lead research projects to formulate stochastic network architecture search problems; develop state-of-the-art learning techniques to expand machine learning for high performance computing and cloud computing, developing network compression algorithms to enable deep learning on edge devices.
- Work in a highly collaborative environment with data scientists, machine learning scientists, remote sensing scientists, engineers, physicists and geographers to deliver systems from prototyping to production level.
- Requires a Ph.D. with a minimum of 6 years relevant experience, a M.S. with a minimum of 12 years experiences, B.S. with 15 years of experience and a focus on computer science, electrical and computer engineering, applied mathematics, machine learning or related fields.
- Experience programming in Python, C, C++ or other production languages, using tools like TensorFlow, PyTorch, Apache Spark.
- Qualified candidates should have demonstrated experience in machine learning for high performance computing and geographic data science related research areas.
- Experience in prototyping and deploying models on GPU accelerated high performance computing platforms.
- Excellent interpersonal skills with a demonstrated leadership ability and a strong commitment to a teaming environment.
- Strong written and oral communication skills.
- Experience working with large spatio-temporal datasets, remote sensing imagery and time series analysis.
- Experience in optimization methods, machine learning, applied mathematics, statistical learning methods.
- Experience with stochastic network architecture search algorithms.
- Good overall knowledge of Linux, OPENMPI, Bash, Csh and Bourne Scripts.
- Hands-on experience with training machine learning models on high performance computing infrastructures with GPU accelerators.
- Demonstrated track record in national and/or international engagements in activities to advance machine learning for high performance computing.
- Experience working for or collaborating with DOE National Laboratories, academic institutions, other research organizations and the private sector.
- Strong drive to learn new topics and skills, to develop innovative products for our customers, and build new programs to expand our capability.