The S. Scott Collis Fellowship in Data Science seeks applicants with a demonstrated background and interest in the interdisciplinary domain of data science to include mathematics, computer science, artificial intelligence, statistics, computer engineering, and related areas. Collis Data Science Fellows work at the junction of cutting-edge data science techniques and Sandia’s national security mission and application space. Examples of relevant data science techniques range from machine learning and deep learning to advanced statistical methods to optimization techniques to methods designed for specialized hardware and beyond. Examples of the scientific and engineering applications relevant to Sandia’s diverse missions include nuclear deterrence and nonproliferation, numerical modeling and scientific simulation, energy systems and power grid, cyber security, operations research, geospatial analysis, neuromorphic computing, and many other national security domains.
The S. Scott Collis Fellowship in Data Science is a prestigious postdoctoral fellowship that provides opportunities for highly motivated researchers to address challenging problems in national security applications. Scott Collis was the director of Sandia’s Center for Computing Research from 2017 to 2022. Due in large part to his recognition of and advocacy for data science as a pivotal research area for impacting problems of national importance, the Sandia Data Science Postdoctoral Fellowship was created in 2021 to attract the best and brightest postdoctoral researchers in data science. The fellowship was renamed the S. Scott Collis Fellowship in Data Science in honor of Scott, who passed away in 2022 after battling cancer. Scott’s vision will continue to be carried out through the innovative research of selected fellows working at the intersection of academic excellence and impactful applications.
As a Collis Fellows, you will pursue a combination of both self-directed research in your topic areas that they select, and integration with existing Sandia R&D projects, all under the guidance of Sandia staff mentors. Fellows are expected to publish the results of their work in leading journals and present research at top-tier conferences. Selected Fellows will receive a two-year appointment with the potential option for a third year, which includes a highly competitive salary, moving expenses, and a generous professional travel allowance.
Sandia National Laboratories is dedicated to nurturing a culture compatible with a broad group of people and perspectives. Consistent with this dedication, we seek applicants from diverse backgrounds, and we foster an inclusive research community.
To be considered for this fellowship, applicants must have completed, or are pursuing, a Ph.D. in a data science related area, such as mathematics, computer science, statistics, or engineering, conferred within the past three years, or anticipated completion of Ph.D. requirements by commencement of appointment. Furthermore, applicants must have appropriate research experience in one or more areas of data science as evidenced by a strong record of research publications and presentations. Applicants should also have no previous postdoctoral appointments at a national laboratory and must be eligible to acquire a DOE security clearance, which requires US Citizenship.
Application process
Applicants should submit a one- or two-page summary, excluding references, of a data science research problem and proposed approach. We are interested in understanding the types of problems, data, and techniques that candidates hope to pursue during a potential fellowship. Insights into how the proposed work may relate to national security problems are welcomed, but not required. If selected for the fellowship, the candidate will work with a Sandia staff mentor to align the proposed work to Sandia’s mission needs and expand the summary into a formal proposal for the funding that will support the next two years of their research at half-time. The other half of the selected fellow’s time will come from an existing Sandia R&D project, giving the fellow the opportunity to collaborate on a research team.
To apply to the S. Scott Collis Fellowship in Data Science, please complete the following steps:
- Read the complete Fellowship job description, including required and desired applicant qualifications.
- Submit a single PDF file containing your cover letter, CV, and a 1-2 page research proposal when applying to Job ID 693857
- Have three letters of recommendation sent to datasciencefellow@sandia.gov with “Collis DS Fellowship” as the subject line.
Applicants who are selected for on-site interviews should be available to travel during the first two weeks of December 2024.
Complete applications received through October 28, 2024 will receive full consideration.
The S. Scott Collis Fellowship in Data Science is supported by Sandia’s Computing and Information Science Research Foundation (CIS RF). The CIS RF advances research and development in computing, mathematics, and information sciences to position Sandia at the frontiers and intersection of science, engineering, and national security. Collis Fellows work on interdisciplinary teams that span the CIS research areas and have access to cutting-edge computational resources.
Current Collis Data Science Fellows
Alejandro Diaz, 2024 Collis Fellow
Quadratic-maniforld reduced-order models for multi-component systems
Alejandro is a S. Scott Collis Fellow in Data Science in the Quantitative Modeling & Software Engineering department at Sandia National Labs. His work focuses on nonlinear model reduction of multi-component systems, and is broadly interested in developing modeling and simulation techniques at the intersection of data science/machine learning and classical numerical methods. He received his MA and PhD in Computational and Applied Mathematics from Rice University and a BS in Mathematics from University of Maryland, College Park. While a PhD student, he was a DoD National Defense Science and Engineering Graduate Fellow and completed a research internship at Lawrence Livermore National Laboratory and a data science internship at Microsoft.
Alejandro’s Fellowship project considers reduced order models (ROMs), which offer an attractive solution to mitigating the high computational costs and lengthy runtimes of large-scale simulations. While numerical simulation allows scientists and engineers to prototype new technologies before entering costly manufacturing and deployment stages, high computation costs and lengthy runtimes make many-query tasks like uncertainty quantification and design optimization infeasible given time and budget constraints. Model reduction alleviates these concerns by systematically extracting the relevant dynamics of the large-scale full-order model (FOM) from simulation data to construct a computationally efficient ROM, which can substitute for the FOM in many-query tasks to significantly reduce the associated cost and runtime without sacrificing accuracy. This project focuses on advancing recently developed quadratic-manifold ROMs, which overcome known deficiencies of linear model reduction approaches, in the context of multi-scale and multi-physics problems.
Esha Datta, 2023 Collis Fellow
Bayesian inference and measures of information for persistent homology
Esha Datta is a S. Scott Collis Data Science Fellow with the Secure Algorithms department. Her research is at the intersection of topological data analysis and machine learning, with particular focus on interpretability, generalizability, and information-theoretic measures of trustworthiness. She received her PhD in Applied Mathematics from University of California, Davis, where she developed a reliable and interpretable topological algorithm for biomedical data analysis. A central theme in her work is the dissemination and application of tools from pure mathematics to real-world problems.
Esha’s Fellowship project investigates the connections between topological data analysis (particularly persistent homology) and Bayesian inference. Persistent homology is a topological tool capable of uncovering shape characteristics in data that may be missed by traditional methods, particularly when data is noisy, irregular, or limited. We aim to leverage information-theoretic connections to Bayesian inference for rigorous uncertainty quantification in persistent homology and develop computational tools for use in high-consequence application areas.
All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability, or veteran status and any other protected class under state or federal law.
Sandia invites you to review the Equal Employment Opportunity posters which include EEO is the Law, EEO is the Law Poster Supplement, and Pay Transparency Nondiscrimination Provision.
Sandia is a drug-free workplace. As a national laboratory funded by a U.S. government agency, we are subject to federal laws regarding illegal drug use. Illegal use of a controlled substance, including marijuana even in places where it does not violate state law, may impact your ability to obtain and/or maintain a Department of Energy security clearance, and may result in the withdrawal of an employment offer or termination of employment.