Postdoctoral researcher
The Sensor Science Division at the National Institute of Standards and Technology (NIST) is seeking a
researcher to advance the application and scientific understanding of Artificial Intelligence and Machine
Learning (AI/ML) in forensic firearm and toolmark analysis. Forensic examiners compare toolmarks on
cartridge cases or bullets to evaluate whether they were fired from the same firearm. A similar
comparison is made with marks from other tools, such as pliers and additive manufacturing systems
(e.g. 3D printers). These analyses are currently subjective and rely heavily on examiner expertise. This
position will lead the development, evaluation, and characterization of AI/ML methods to improve the
objectivity, reproducibility, and accuracy of toolmark pattern evidence analysis. The researcher will
address critical challenges in the application of AI/ML to forensic science, including transparency,
robustness, and bias, through the development of guidelines for training and validation datasets;
procedures to rigorously characterize model performance, uncertainty, and operational limitations; and
approaches to provide insight into model decision-making processes. The researcher will collaborate
with interdisciplinary teams and communicate findings through technical reports, publications, and
presentations, contributing to the development of scientifically grounded and standards-based
approaches for forensic evidence evaluation.
Artificial Intelligence for Forensic Firearm and Toolmark Analysis
- Ph.D. or master’s degree in computer science, physics, engineering, statistics, forensic science,
or a closely related field. - Research experience in the development and application of AI/ML for image analysis.
- Proficiency with Python or MATLAB.
- Evidence of independent research experience and a strong enthusiasm for learning new
theoretical, computational, and experimental techniques. - Strong oral and written communication skills.
Responsibilities include but are not limited to:
- Develop a forward-looking research program on AI/ML in forensic firearm and toolmark analysis
and collaborate on cross-cutting techniques for other types of pattern evidence. - Lead the development of a computational pipeline for the consistent segmentation of toolmark
images using AI/ML methods. - Investigate application of AI/ML to address major challenges in forensic toolmark analysis, such
as characterizing toolmark quality and, ultimately, the direct comparison of toolmark images. - Benchmark AI/ML-assisted results against those obtained with procedural algorithms and
traditional human examinations. Investigate where AI/ML can provide the most significant
improvements in objectivity and efficiency. - Investigate approaches for quantifying the uncertainty of AI/ML outputs to ensure objective
communication of the strength of the evidence in courtroom testimony. Explore the feasibility
of explainable AI (XAI) frameworks to inform examiners on the toolmark features that drive the
model’s output. - Collaborate with the forensic community to translate research findings into actionable
standards and best practice guides. - Publish research findings in peer-reviewed journals and present results at scientific conferences.