MACHINE LEARNING RESEARCH SCIENTIST

Req #: 189894
Department: CICOES
Posting Date: 05/05/2021
Closing Info: Closes On 05/12/2021
Salary: Salary is commensurate with education and experience.
Limited Recruitment: Other
Limited Recruitment Other: OPEN TO UW EMPLOYEES ONLY
Shift: First Shift
As a UW employee, you have a unique opportunity to change lives on our campuses, in our state and around the world. UW employees offer their boundless energy, creative problem solving skills and dedication to build stronger minds and a healthier world. UW faculty and staff also enjoy outstanding benefits, professional growth opportunities and unique resources in an environment noted for diversity, intellectual excitement, artistic pursuits and natural beauty. The University of Washington (UW) is proud to be one of the Nation's premier educational and research institutions. Our people are the most important asset in our pursuit of achieving excellence in education, research, and community service. UW is in the greater Seattle metropolitan area, with a dynamic, multicultural community of 3.7 million people and a range of natural environments from mountains to ocean. The UW is a community of 80,000 students, faculty and staff including 25% first-generation college students, over 25% Pell Grant students and faculty from over 70 countries. The Cooperative Institute for Climate, Ocean, and Ecosystem Studies (CICOES) has existed since 1977 for the purpose of fostering research collaboration between UW and the National Oceanic and Atmospheric Administration (NOAA). CICOES's research is at the forefront of investigations on climate change, ocean acidification, fisheries assessments, and tsunami forecasting. The impact of CICOES's environmental research is felt by communities all over the world, and a broad variety of perspectives and life experiences is essential to the success of this research. We encourage candidates from groups historically and currently underrepresented in this field to apply. Please read about our commitment to diversity, equity, and inclusion here: https://cicoes.uw.edu/about/diversity/. CICOES has an exciting opportunity for a Machine Learning Research Scientist. This Machine Learning Research Scientist conducts research in support of the CICOES project Abundance, trends, and distribution of Alaska seals in Arctic and sub-Arctic marine ecosystems, in cooperation with the Alaska Fisheries Science Center's Marine Mammal Laboratory (MML). The primary research activities are: 1) Testing, evaluation, and refinement of methods for acquiring and processing digital images obtained in instrument-based surveys of pinnipeds, and 2) Strengthening field and analytical capabilities for employing unoccupied aircraft systems (UAS) to monitor pinnipeds in habitats that are particularly remote and costly to monitor from traditional, occupied aircraft. This project was designed to improve understanding of the status, and factors responsible for the dynamics of Arctic and sub-Arctic pinniped populations; the information is key to understanding their roles in marine ecosystems, and human impacts on them from fisheries and climate disruption. This position independently and with minimal supervision, evaluates, selects and applies machine learning approaches and techniques in support of the project and develops models for detection and classification of animals in aerial imagery as follows: •Develop machine learning (ML) models for detection and classification of animals in nadir and oblique aerial imagery; •Advise MML AI practitioners on effective workflows and data structures for managing ML projects based on large image and acoustic spectrogram databases; •Build and maintain annotation databases to interface with primary SQL databases; •Work with MML scientists to identify appropriate image sets for training and testing ML models; •Ensure models function within the Kitware VIAME pipeline, or provide interface for viewing and evaluating model performance; •Explore approaches for internal ML development and identify projects/datasets appropriate for AI specialist partners through UW, NASA, and private industry; •Identify avenues to increase efficiency of annotation process (multi-model approach, outsource options, etc.) while maintaining high scientific standards; •Work with MML scientists to develop and apply methods for evaluating the performance of target species detection and classification processes (e.g., detection rates, misclassification rates, false positive rates); •Participate in aerial surveys by operating remote sensing equipment and managing the collection of images and related data; •Produce summary statistics and other graphics to aid in communication of detection and classification; •Work with MML scientists to write reports and peer-reviewed manuscripts, and prepare presentations; and, •Collaborate with researchers and partners within and outside of NOAA and UW to facilitate communication with AI specialist groups working with NOAA data. As a UW employee, you will enjoy generous benefits and work/life programs. For a complete description of our benefits for this position, please visit our website, click here. (https://hr.uw.edu/benefits/wp-content/uploads/sites/3/2018/02/benefits-professional-staff-librarians-academi-staff-20210208.pdf) Required Qualifications: •A Bachelor's degree in Computer Science or related field •At least 2 years of related work experience •Training and experience in machine learning (ML) and deep learning techniques •Demonstrated experience training models using machine learning and deep learning algorithms to detect and classify marine mammals in aerial imagery •Programming experience in Python, C/C++, and other languages commonly used for software development and machine learning/artificial intelligence •Background in machine learning frameworks such as Darknet, TensorFlow, or Keras, and experience with implementing various object detection algorithms (e.g., YOLO, CNN, RCNN, etc.) •Proficiency with image labeling/annotation for ML model training and annotation database management •Ability to communicate with research teams to understand the goals of individual projects and develop plans to improve image/data processing approach •Willingness to participate in fieldwork including flying in small aircraft in remote locations •Experience with full-stack software development (i.e., develop user interfaces and underlying software infrastructure •Note: equivalent experience may be substituted for degree requirement Equivalent education/experience will substitute for all minimum qualifications except where there are legal requirements such as license/certification/registration. Desired Qualifications: •Experience applying Kitware VIAME computer vision tools and pipeline frameworks to large datasets of high-resolution images •Experience with developing software and algorithms for use on cloud computing platforms (e.g., Amazon Web Services (AWS), Microsoft Azure, Google Cloud) •Experience with developing and deploying containerized applications •Experience with interfacing with hardware, such as cameras, sensors, or other physical devices using APIs or other frameworks Application Process: The application process for UW positions may include completion of a variety of online assessments to obtain additional information that will be used in the evaluation process. These assessments may include Workforce Authorization, Cover Letter and/or others. Any assessments that you need to complete will appear on your screen as soon as you select "Apply to this position". Once you begin an assessment, it must be completed at that time; if you do not complete the assessment you will be prompted to do so the next time you access your "My Jobs" page. If you select to take it later, it will appear on your "My Jobs" page to take when you are ready. Please note that your application will not be reviewed, and you will not be considered for this position until all required assessments have been completed.
University of Washington is an affirmative action and equal opportunity employer. All qualified applicants will receive consideration for employment without regard to, among other things, race, religion, color, national origin, sexual orientation, gender identity, sex, age, protected veteran or disabled status, or genetic information.