Collaborative research: research initiation: forming t-shaped civil systems engineers in western pennsylvania via AI-enabled communication coach for large-scale infrastructure

The National Science Foundation Division of Engineering Education and Centers awarded Carnegie Mellon University $150,000 on January 1, 2027, under the Engineering program (CFDA 47.041) to develop and test an AI-enabled communication coach for civil engineering teams working on large-scale infrastructure problems. The award funds research and educational development through December 31, 2028, with place of performance in Pittsburgh, Pennsylvania.

The project will create a computer system that observes engineering teams discussing realistic infrastructure scenarios in a simulated environment and delivers personalized feedback to improve cross-disciplinary communication and coordination. The initiative targets the formation of T-shaped engineers—professionals combining deep technical expertise with broad communication and teamwork skills—in collaboration with construction industry partners and a municipal public works department in western Pennsylvania. The project simultaneously trains two early-career engineering faculty members in rigorous education research methods, directly supporting the NSF Professional Formation of Engineers: Research Initiation in Engineering Formation (PFE: RIEF) program. This assistance is structured as a project grant.

Abstract

The roads, bridges, and water systems that americans rely on every day are aging, and rebuilding them requires large teams of engineers, planners, contractors, and equipment operators who must work together under tight budgets, schedules, and safety requirements. When these professionals fail to communicate effectively, such as when a designer's intent is not clearly conveyed to a construction crew, or when a field observation is not accurately reported back to the office, the results can include safety incidents, costly rework, and schedule delays that ultimately cost taxpayers billions of dollars each year. Yet most civil engineers receive little formal training in how to communicate across disciplines; they are expected to learn these skills informally, over many years on the job. This project will address that GAP by developing and testing an artificial intelligence (AI)-enabled ?communication coach?: a computer system that will observe how engineering teams talk through realistic infrastructure problems in a simulated environment and will give them personalized feedback on how to communicate more clearly and coordinate more effectively. Working with construction industry partners and a municipal public works department in western pennsylvania, the project will prepare ?t-shaped? Engineers, professionals who combine deep technical expertise with the broad communication and teamwork skills needed to lead complex infrastructure projects. By simultaneously training two early-career engineering faculty members in rigorous education research methods, the project will directly advance the goals of the national science foundation?s professional formation of engineers: research initiation in engineering formation (PFE: rief) program, and IT will strengthen the safety, efficiency, and resilience of the nation?s infrastructure workforce. This project will establish a scientific foundation for forming t-shaped civil systems engineers by investigating three research questions: (1) what balance of disciplinary depth and cross-disciplinary breadth will define t-shaped competence; (2) how an AI-enabled communication coach will scaffold the development of engineers? Communication and coordination capabilities; and (3) how formative and summative metrics will capture the growth and impact of engineering communication performance in authentic contexts. The research will proceed through two research thrusts and one evaluation thrust. Thrust 1 will characterize the t-shaped knowledge space for large-scale highway construction by building a cross-domain conceptual ontology from academic curricula, vocational training materials, and structured expert interviews; developing a disciplinary knowledge index and scenario-based simulation instruments; and conducting an observational study with 12?15 three-person teams to model how individual depth and breadth profiles predict team communication performance. Thrust 2 will design and implement the AI-enabled communication coach, which will integrate a multimodal natural-language-processing pipeline that codes team dialogue using the problem?cause?solution?outcome (pcso) framework, personalized cognitive models of each learner built on instance-based learning theory, and a coaching engine that compares learner and expert models to deliver explainable, adaptive feedback during and after simulation sessions. The evaluation thrust will assess the coach?s effectiveness through a randomized pre-test/post-test experiment with 20?26 teams of students and professionals, a formative analysis of collaborative learning trajectories, and external validation by learning-science experts and an industry advisory board. The project will be conducted through a partnership between carnegie mellon university and the university of pittsburgh, with authentic scenarios, field data, and participants provided by the constructors association of western pennsylvania and cranberry township public works. Expected outcomes will include an empirically validated model of t-shaped engineering competence, an open-source AI coaching platform, validated formative and summative assessment instruments for engineering communication, and durable engineering-education research capacity for two new investigators. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.