Hello Aaron and Oscar Sir, 
I am a contributor, but I had an idea which could be used as a GSoC 
project. As per as my knowledge this feature doesn't exist yet in SymPy, so 
I would like some validation and feedback from your end.
*A robust hand written Mathematical symbolic expression pipeline, with 
formal verification, resolving for any ambiguity, and SymPy native parsing 
and evaluation.*
Input is taken as a hand written math expression (image) and then output is 
given in form of SymPy Code or evaluated output.
Further, this can be implemented with a LLM to provide AI based insights on 
the expression  - if its accurate, logical or etc.

Please do provide your feedback for this idea. And, I am looking forward to 
working with you guys for this year's GSoC!

Thank You,
Arka Saha
On Friday, 23 January 2026 at 07:21:01 UTC+5:30 Oscar wrote:

> On Thu, 22 Jan 2026 at 17:48, Aaron Meurer <[email protected]> wrote:
> > My
> > suggestion would be to have a policy that LLMs are not allowed to be
> > used when writing the application, except for use for
> > translation/cleaning up the grammar of the text.
>
> I don't think that we should list any exceptions like that. The
> reality is that if I am ranking applications then anything that
> remotely looks like the output of an LLM will be straight in the bin.
> It might be that they did not use an LLM but just wrote something that
> resembles an LLM: it will still be in the bin. I don't think we should
> allow anyone to be confused about that.
>
> --
> Oscar
>

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/sympy/bafdff2f-d0fb-4a1f-ab22-978c50c6841en%40googlegroups.com.

Reply via email to