Recent graduate [1] broadly interested in optimization, machine learning, and scientific computing, mainly using Python, C++, and C. In particular, most of my personal and professional work has been done in Python, C++17, and C++20.

For fun, here's a toy norm-constrained convex optimization problem and a plot of its solution against the objective's minimum. The Python script used to solve the problem and generate the plot can be found in my profile repository.

npyglue

C++ project aimed at simplifying creation of NumPy arrays from C++ objects providing the backing memory. Supported backing objects include STL vectors, Armadillo and Eigen matrices, and PyTorch C++ CPU/GPU tensors. Integration layers include the Python C/C++ API layer and SWIG. Unit and integration tests in C++/Python are all orchestrated with CTest.

numpy-lapacke-demo

Python C extension implementations of linear regression using QR/SVD and Newton's method with diagonal Hessian modification using the Python C API and NumPy C API to work with Python objects on the C level. Computations are done using CBLAS/LAPACKE routines operating directly on NumPy array memory. All public and private methods are rigorously unit tested using pytest.