Lead, Follow, or Get Out of the Way - Julia and Threaded Python

Explore the complex interoperability challenges between Julia and Python's threading models in this 28-minute conference talk from JuliaCon Global 2025. Discover why Python's threading architecture creates performance bottlenecks when wrapping Julia packages, despite Julia's goal of solving the "two language problem" that typically requires high-performance languages like C, C++, or FORTRAN underneath Python interfaces. Learn about the technical obstacles that arise when institutional reluctance to adopt Julia drives demand for Python wrappers around Julia packages, and understand how Python's Global Interpreter Lock and threading limitations can severely impact the performance of threaded Julia code. Examine practical workarounds and emerging solutions in Python's ecosystem, while gaining insights into current strategies developers can implement when creating Python wrappers for high-performance Julia applications. Delve into potential improvements for community-controlled tools like PythonCall.jl and juliacall that could streamline the development process for wrapper creators, and participate in discussions about advancing cross-language interoperability between these two important scientific computing ecosystems.