CIRES/ EARTH LAB Machine Learning Applications to Earth Observations of Natural Hazards Post-Doc

Job Summary

Earth Lab, funded by the University of Colorado Boulder’s “Grand Challenge: Our Space, Our Future” and part of CIRES, seeks post-doctoral researchers to join a dynamic team pushing the frontiers of coupled Earth and social system science (http://www.colorado.edu/earthlab/). Earth Lab’s mission is to harness the data revolution through research, analytics, and education to accelerate understanding of global environmental change to help society better manage and adapt.

Earth Lab is seeking a Post-Doctoral Research Scholar to lead a research agenda in the following area: Machine Learning Applications to Earth Observations of Natural Hazards. This targeted research area represents Earth Lab’s efforts to explore natural and social system vulnerability and resilience to global environmental change, while also capitalizing of the diversity of data available to generate new insights. All postdoc positions are for one year with the possibility of extension based on performance and funding availability.

Earth Lab seeks a Postdoctoral Research Scholar to advance machine learning applications in analyzing and integrating Earth observations, and non-traditional data sources - like Twitter or Zillow - to better understand global change and the consequences to ecosystems and society. This candidate will push forward on research frontiers that are emerging as methodological tools from the deep learning revolution spillover into the Earth and environmental sciences. These include applications of generative models, generative adversarial networks, self-supervised or semi-supervised learning for sparse data, transfer learning, reinforcement learning, and science-based deep learning that combines dynamical or physical models with neural networks. The candidate chosen for this position would have flexibility in terms of the particular systems, questions, and machine learning methods that are studied. Example research topics include:

• Benefits of super-resolution techniques applied to satellite imagery assessed by classification accuracy, unmixing ratios, and/or physical process modeling.
• Combining co-located point cloud data (e.g., structure from motion) with RGB, multispectral and thermal imagery for surface characterization.
• Integration of commercial satellite imagery with long-term NASA and USGS research satellite observations (off-nadir view angles, inconsistent radiometry)
• Real-time disaster response support that integrates disparate streams of observations including remotely sensed and social media data.
• Science-based deep learning that permits inference about dynamical models and facilitates prediction.
• Reinforcement learning to evaluate optimal control of wildfires.
• Self-supervised learning for plant classification from hyperspectral imagery.
• Scalable Bayesian deep learning for earth science applications that accommodates parameter uncertainty (e.g., variational inference).

The research goal of this work is to advance our understanding of some aspect of:

• Predictive inference in support of real time, short term, and/or long term prediction.
• Object detection, segmentation, or classification of remotely sensed observations that leverages convolutional neural networks with limited training data.
• Development of key Earth science training datasets to support future work.
• Use of cloud computing resources to scale analytics.
• Natural hazards, extreme events, and global change using multiple streams of information.

The University of Colorado Boulder is committed to building a culturally diverse community of faculty, staff, and students dedicated to contributing to an inclusive campus environment. We are an Equal Opportunity employer, including veterans and individuals with disabilities.

Who We Are

At CIRES, the Cooperative Institute for Research in Environmental Sciences, more than 800 environmental scientists work to understand the dynamic Earth system, including people’s relationship with the planet. CIRES is a partnership of NOAA and the University of Colorado Boulder, and our areas of expertise include weather and climate, changes at the Earth’s poles, air quality and atmospheric chemistry, water resources, and solid Earth sciences. Our vision is to be instrumental in ensuring a sustainable future environment by advancing scientific and societal understanding of the Earth system.

https://cires.colorado.edu/

Earth Lab, funded by the University of Colorado Boulder’s “Grand Challenge: Our Space, Our Future” and part of CIRES, seeks post-doctoral researchers to join a dynamic team pushing the frontiers of coupled Earth and social system science ( http://www.colorado.edu/earthlab/ ). Earth Lab’s mission is to harness the data revolution through research, analytics, and education to accelerate understanding of global environmental change to help society better manage and adapt.

What Your Key Responsibilities Will Be

• Work with individual team to meet research goals of this position, with the expectation of submitting one manuscript for publication to a peer-reviewed journal by the end of Year 1, and contribute to and/or submit an independent funding proposal.
• Use research agenda to develop use cases for the Analytics Hub (staff and Viz-studio) with heterogeneous and/or big data streams that require specialized data management, analysis, and/or visualization support (with emphasis on data from the Earth Observation enterprise, from ground-based to space-based sensors).
• Contribute to discussions with industry, federal, and other academic partners, and contribute to opportunities to make Earth Lab visible in the community.
• Contribute to the open, reproducible science objectives of Earth Lab. This could include contributing or publishing well-documented data sets or code recipes that could serve multiple users within Earth Lab or other audiences.
• Work on synergistic activities across the post-doc cohort and Science Teams. This could include collaboration on external proposals, papers, workshops, opportunities with industry/federal partners, or other collaborative activity and should amount to 20% effort.
  

What You Should Know

• This position is located at the University of Colorado at Boulder's main campus.
• All postdoc positions are for one year with the possibility of extension based on performance and funding availability.
• To learn more about Earth Lab, visit Earth Lab’s website ( http://www.colorado.edu/earthlab/ ) and Earth Lab’s learning portal ( www.earthdatascience.org ).
• Review of applications will begin October 18. The position will remain open until filled.

What We Can Offer

• We can offer a competitive salary and a comprehensive benefits package.
• The hiring annual salary range for this position is $55,000 - $66,000.
• There will be no relocation.

Benefits

The University of Colorado offers excellent benefits , including medical, dental, retirement, paid time off, tuition benefit and ECO Pass. The University of Colorado Boulder is one of the largest employers in Boulder County and offers an inspiring higher education environment. Learn more about the University of Colorado Boulder .

Be Statements
Be Driven. Be Respectful. Be Boulder.

What We Require

• Ph.D. in a related field is required, such as applied mathematics, computer science, geography, ecology, environmental studies, remote sensing, or other.
• Background expertise in machine learning, deep learning, artificial intelligence, and/or statistics possibly with applications to environmental data.
• Applicant must have demonstrated interest and skills in one or more of the approaches described above (e.g., machine learning approaches, remote sensing, data integration across multiple sources, etc.).
• A strong quantitative background is necessary.
• The ability to work as part of an interdisciplinary team.

What You Will Need

• Experience in, or willingness to learn, appropriate programming and data analytic tools. Ideally the candidates will have experience in programming languages (e.g., R, Python, or others), can work in different environments (e.g., Linux), and are well versed in geospatial analysis software (e.g., QGIS).
• Demonstrated contributions to open science (i.e., publicly available and/or reproducible data, code, workflows, and/or tools) or willingness to contribute to open science.
• Experience in integrating and analyzing large, and/or heterogeneous datasets.
• Experience in working with a high performance computing or cloud computing environment is a plus (Earth Lab supports both HPC & cloud compute on AWS).
• Demonstrated publication and grant-writing skills.
• Team spirit and interest in interdisciplinary settings, with a willingness to engage with Earth Lab’s Analytics Hub and Education Initiative teams.

Special Instructions

To apply, please submit the following materials:

1. Resume or CV
2. Cover letter addressed to the Search Committee briefly describing your qualifications, professional goals, and specific interest in this position.
3. List of contact information for 3 references who will be willing to write a confidential Letter of Recommendation for you.

If you are selected as the finalist, your degree will be verified by the CU Boulder Campus Human Resources department using an approved online vendor. If your degree was obtained outside of the United States, please submit a translated version as an optional attachment.

Review of applications will begin October 18. The position will remain open until filled.

Note: Application materials will not be accepted via email. For consideration, applications must be submitted through CU Boulder Jobs .

Posting Contact Information

Posting Contact Name: Maxwell Joseph

Posting Contact Email: Maxwell.b.joseph@colorado.edu