The Duke Master in Interdisciplinary Data Science (MIDS) program together with the newly merged Duke University Energy Initiative and Nicholas Institute for Environmental Policy Solutions are seeking a Postdoctoral Associate who (1) will make significant contributions to the research agenda of the Energy Initiative/Nicholas Institute (2) will work to mentor, train, and educate MIDS students. The home department for the associate will be the Department of Electrical and Computer Engineering.
Research in the Energy Initiative/Nicholas Institute relevant to this position is focused on advancing accessible, affordable, reliable, and clean energy systems. Specifically, the individual will work in the Energy Data Analytics Lab, which develops and applies advanced data analytics tools to transform diverse energy data into insights that lead to energy system performance improvements. The nexus of energy systems, climate change, and societal impacts presents an interconnected web forming a wicked problem; the Energy Data Analytics Lab seeks to increase the benefits of energy systems to society, decrease the associated climate risks, and increase the equity of both benefits and costs across society—nationally and globally.
The mission of the Energy Data Analytics Lab is to aid these efforts in three ways: by developing (a) objective evidence (data), (b) tools for gathering those data at scale, and (c) analytic tools for enabling effective decision-making and data literacy.
The Lab will advance both the technological and application innovations required to bring about a more sustainable future through our energy system planning and operations. In that process the Lab engages MIDS students in that research.
Education in the MIDS Program culminates in a year-long intensive capstone project, in which teams of MIDS students make substantial contributions to real, complex projects between non-academic partners and Duke researchers who have interests and domain expertise aligned with partners’ goals. Capstones include research projects both with a variety of industrial partners and with the Energy Initiative and partners of the Energy Initiative. A major goal for the capstones is to help transform the students from technical experts to proactive researchers who are capable of designing, refining, and executing research agendas and also who are able to clearly communicate and disseminate their findings.
The educational component of this position involves (a) facilitating and designing these capstone projects, (b) providing mentorship to some of the MIDS students, and (c) helping train the MIDS students to become proficient and proactive researchers.
There is overlap between the research and educational components of these positions. Individuals will have opportunities to create capstone projects related to their interest and research. Individuals will also assist with existing capstone projects that serve to advance the research of the Energy Initiative/Nicholas Institute. Finally, individuals will have an opportunity to recruit and interface with industrial partners, government agencies, NGOs, and more to help generate, refine, and resource problems and projects that are of direct value to these agencies. Although there will be some harmony between the research and education efforts, work will be divided up about equally between these two components.
Work Performed Research Activities The individual will conduct research on signature projects of the Energy Data Analytics Lab. One such project will investigate how remotely sensed data (satellite imagery, drone imagery, lights at night data, etc.) can be analyzed to produce valuable insights about energy systems, resources, and their societal impacts. The individual will explore the productive applications of existing machine learning techniques to facilitate assessment of infrastructure such as solar photovoltaics, power plants, transmission networks, buildings, and their associated energy consumption/generation; they will identify regions of the world with limited access to electricity to develop pathways towards electrification; etc. They will develop and evaluate new machine learning methodologies for extracting insights from remotely sensed data such as novel deep learning architectures, self-supervised learning and foundation models, etc.
From the above work, the individual will produce written research reports including journal publications; disseminate research through oral presentations including conference presentations and workshops and help to organize relevant research workshops hosted by Duke and its partners. They will design and publish web-based media such as interactive data visualizations, geospatial data visualizations, and the websites on which they reside. They will manage GPU servers and associated data repositories. They will develop novel machine learning training datasets especially remotely sensed satellite, aerial, and drone imagery.
In addition, they will participate in additional research projects as needed. Past examples include investigating the integration of intermittent resources into the grid using statistical modeling and data analysis techniques; automated building energy assessment through non-intrusive load monitoring (the disaggregation of whole-building energy consumption into appliance-level information).
Educational and Outreach Activities Capstone Mentorship. The individual will mentor teams of data science master’s students as they work on their year long Capstone projects. These projects pair student teams with external partners to generate insights and recommendations unique to the partner’s needs. Mentors must be able to regularly meet with the students to provide guidance, track progress, provide written feedback on papers, and resolve potential personal conflicts between team members and partners. Mentors will need to occasionally meet with project stakeholders to help students interpret and understand stakeholder needs. Mentors will report to the program director and will keep them updated with student progress. Good communication and project management skills are required. Instructor for Elective Course. The individual will design and teach an elective course. This course will either be a 1.5 credit (half-credit) course taught by the individual or a co-taught 3 credit (full-credit) course jointly taught by the individual and a collaborator. This course will be offered as an elective in the Department of Electrical and Computer Engineering and cross-listed in the MIDS program and will be relevant to the individual’s expertise and instructional needs of the MIDS and Electrical and Computer Engineering program. Possible topics include, but are not limited to, geospatial data science and machine learning, applied computer vision, and advanced data visualization.
The individual will also assist with the broader activities of the Energy Initiative/Nicholas Institute, as needed, including developing public-facing blog posts; assisting with the development and implementation of educational modules for classroom and laboratory use relevant to ongoing research and/or MIDS educational priorities; student mentoring with MIDS students as well as Data+ and Bass Connections student research programs; and other duties as assigned. In addition, the individual will assist in recruiting partners and developing projects for each following year’s capstone projects.