Not quite a positive (it's ready now!) answer, but there's some interesting work on denoting problems, and constructing numerical methods for systems like the o

Not quite a positive (it's ready now!) answer, but there's some interesting work on denoting problems, and constructing numerical methods for systems like the one you're describing -- I believe the design of this library (while not yet mature) would support the workflow you described (including both analytic and numerical solutions): https://github.com/lecopivo/SciLean https://github.com/lecopivo/SciLean To your last point, the idea is that numerical approximations can be introduced (and introduction will ask for proofs of validity! but you can ignore "the proving" via `sorry`) to go from un-executable math notation (in Lean4) to executable! Whether the proof goes through doesn't affect the final executable.