Python is programming
language. One formulates ideas in spoken language and then translates them into Python. Can you spell it out - what is your idea how to reach objective?
If you really "Trying to learn and understand loops and iterations" then maybe following can help you.
We have matrix and want to sum sublists.
First of all (if we are not very good at reading squared brackets) we find out what is the structure of our matrix ('for every list in matrix print list/print every list in matrix'):
>>> matrix = [[[4,3], [3,1], [2,0]], [[0,1], [1,2], [3,3]]]
>>> for lst in matrix: # 'for every list in matrix'
... print(lst) # 'display list'
...
[[4, 3], [3, 1], [2, 0]]
[[0, 1], [1, 2], [3, 3]]
We can observe that we have two lists. We need to get 'into' sublists of said lists. For that we add another loop ('for every list in matrix print every sublist in list/print every sublist in list in matrix'):
>>> for lst in matrix: # 'for every list in matrix'
... for sublst in lst: # 'for every sublist in list'
... print(sublst) # 'display sublist'
...
[4, 3]
[3, 1]
[2, 0]
[0, 1]
[1, 2]
[3, 3]
We have 'reached' the needed sublists. But we actually don't want sublists themselves, we want sums. So we can instead sublists print out their sums:
>>> for lst in matrix: # 'for every list in matrix
... for sublst in lst: # 'for every sublist in list
... print(sum(sublst)) # 'display sum of sublist elements'
...
7
4
2
1
3
6
We have needed data, but not in needed datastructure. We have to construct one. We need 'for every list in matrix have list of sums of sublists'. We translate it into Python:
>>> result = []
>>> for lst in matrix:
... result.append([]) # 'for every list in matrix have list in results'
... for sublst in lst: # 'for every sublist in list'
... result[-1].append(sum(sublst)) # 'append sum of sublist to last list in results'
...
>>> result
[[7, 4, 2], [1, 3, 6]]
Using list comprehension this can be expressed as oneliner:
>>> [[sum(sublst) for sublst in lst] for lst in matrix]
[[7, 4, 2], [1, 3, 6]]
I suggest to have a articulated plan and then translate it into Python step by step.