My current research focuses on the applied use of artificial intelligence in real-world settings. My main interest is in how agentic and LLM-based systems can be applied to real-world problems, with attention to reliability, transparency, and responsible use. I am especially interested in bringing rigorous AI and computational methods to domains that have long relied on conventional analytical approaches.
This direction builds directly on a foundation in advanced computational and simulation-based methods. Over the past several years, my research has centered on system dynamics, agent-based modeling, social network analysis, and causal inference. I have used these methods to study complex systems and to support analysis in cases where traditional parametric and regression-based methods fall short. This grounding shapes how I approach new problems and how I design and evaluate emerging AI systems.
Publications
Shrivastav, M. (2026). Talent feeders and U.S. AI innovation capacity: A three-clock sequencing framework for high-skilled immigration, domestic STEM production, and federal R&D (SSRN Working Paper No. 6902640). https://doi.org/10.2139/ssrn.6902640
Shrivastav, M. (2025). Postsecondary enrollment demand-gap estimation model: Projecting the postsecondary demand-gap for California by 2035 [Doctoral dissertation, Claremont Graduate University]. Scholarship @ Claremont. https://scholarship.claremont.edu/cgu_etd/968/
Immormino, J., Hwang, Y., Shrivastav, M., & Webster, K. (2021). Negotiating an inefficient market: An agent-based model approach to property insurance claim negotiations. In D. Cassenti, S. Scataglini, S. Rajulu, & J. Wright (Eds.), Advances in simulation and digital human modeling (Advances in Intelligent Systems and Computing, Vol. 1206, pp. 65–71). Springer. https://doi.org/10.1007/978-3-030-51064-0_9
