the house back to the hallway
the library

Research

Applied math — communication cost and load balancing in matrix multiplication algorithms (Strassen, AlphaTensor-derived methods), with an eye toward parallel and distributed systems.

papers & working drafts
Applied MathPhD Research

Communication-Efficient Matrix Multiplication via Modified Strassen Schedules

Working paper on reducing inter-processor communication cost in Strassen-derived algorithms across distributed compute nodes. Abstract and PDF link go here.

Stony Brook University — Applied Mathematics
Parallel ComputingMPI

Load Balancing for Tile-Cyclic Data Distributions

Coursework-turned-research exploring tile cyclic distribution schemes and their effect on load balance in parallel matrix computation.

Course project → extended writeup
talks & notes
Notes

Reading Notes: AlphaTensor & Fast Matrix Multiplication

Plain-language summary of AlphaTensor-derived algorithms and what they mean for practical communication cost — written for people outside the field.

Long-form note