Armin Gerami

My research interests lie in scientific ML, differentiable programming, and high performance computing.

I'm well familiar with foundations of ML, calculus, linear algebra, probability theory and signal processing. I'm proficient with Python (e.g. implementing/training Transformers), C++ (e.g. implementing digital filters), CUDA (e.g. writing custom gradients in a neural network pipeline), and Verilog (e.g. an accelerator for sparse matrix-vector multiplication).