Saman Amarasinghe

Professor Saman Amarasinghe is a leading academic in the MIT Electrical Engineering and Computer Science Department, specializing in compiler design, high-performance computing, and programming languages. His research bridges theoretical computer science with practical applications, focusing on optimizing compilers for modern hardware architectures. He holds a professorship within the School of Engineering at MIT.

His research interests span artificial intelligence, machine learning integration into compilers, sparse tensor algebra optimizations, and domain-specific languages (DSLs) for high-performance computing. He has pioneered frameworks like TACO (Tensor Algebra Compiler) and GraphIt, which enable efficient handling of complex data structures across diverse computational domains.

Recent work emphasizes compiler-driven optimizations for sparse data formats, machine learning models for code optimization, and unified interfaces for large language models. His projects often address scalability challenges in exascale computing and bioinformatics applications.

Publications highlight advancements in compiler techniques for sparse tensor operations, graph analytics, and memory-efficient algorithms. His contributions to Halide, Simit, and Weld further demonstrate his impact on DSLs for image processing and physical simulation.

Awards and recognitions are integral to his career, though specific personal accolades are not detailed here. His work is characterized by a strong focus on practical compiler solutions for emerging computational challenges.