Hi Brian Simulator Team,
My name is Siddhant, and I am a software engineering student very interested in the project to package Brian’s unit system as a standalone library for GSoC 2026.
I have been exploring brian2/units/fundamentalunits.py and I’m fascinated by how Brian handles dimensional analysis with NumPy. I’m particularly interested in the challenge of decoupling this from the core while maintaining performance and NumPy 2.0 compatibility.
My Background:
I have previously contributed( raised 2 pr’s) to INCF #52( docs: fix incorrect Windows venv activation path #52 by ulekarsiddhant0-boop · Pull Request #66 · INCF/knowledge-space-agent · GitHub ) and #63( test: implement unit tests for knowledge-space api logic #63 by ulekarsiddhant0-boop · Pull Request #65 · INCF/knowledge-space-agent · GitHub )so, I am familiar with the workflow.
I have also contributed to the Apache Software Foundation (2 PRs), where I gained experience working with large-scale codebase infrastructure and rigorous CI/CD standards.
I am comfortable with modern Python packaging (pyproject.toml, setuptools) and unit testing with pytest.
My Initial Thoughts: I’ve started looking at how we might handle interoperability with other libraries like Pint. Should the primary goal be making Brian’s units “convertible” to Pint, or is the focus purely on extracting the current SI-base-unit logic into a dependency-free package first?
I’ve already set up my local environment and passed the unit tests. I’d love to know if there are any “good first issues” or specific refactoring tasks you’d recommend I look at to better understand the current unit-coupling.
Best regards,
ulekarsiddhant0-boop