Johnnie McGray developed and optimized advanced tensor operations in the numpy/numpy repository, focusing on enhancing the einsum function with batched matrix multiplication and improved parsing logic. Using Python and leveraging numpy’s numerical computing capabilities, Johnnie implemented efficient algorithms for high-dimensional tensor contractions, reducing compute time and expanding einsum’s expressiveness. He addressed parsing edge cases, improved documentation, and expanded test coverage to ensure reliability and correctness. Additionally, Johnnie contributed to conda-forge/staged-recipes by creating the initial symmray package recipe, managing dependencies and metadata with YAML, and enabling broader ecosystem adoption through reproducible packaging and dynamic Python versioning.