Making

I work at the intersection of graphs and reasoning — ranking and recommendation, search, graph neural networks, and the context systems that let a model reason over structure instead of flat text.

I’ve been the subject-matter expert on graph databases in both a Fortune 500 bank and an early-stage startup. What I’m working toward is patentable method work in that space, particularly in reasoning and context management.

I published two papers in 2026 during my MS in Computer Science and Engineering at the University at Buffalo. As a graduate researcher at ADAMS Lab, my thesis applied graph-based learning to order-preserving search and recommendation over combinatorial choices — mixed-integer and mixed combinatorial nonlinear optimization, where the space is discrete, non-convex, and large enough that solvers routinely stall short of a good answer. Our approach guides them toward globally feasible solutions faster; how much faster depends on the problem.