I moved each "j" to a position immediately after the imaginary part of the complex number and squeezed out all the spaces to get an example file like this.
(-353.10438+1.72317617j),(-23.16000+0.72512251j)
(-353.10438+1.72317617j),(-23.16000+0.72512251j)
(-353.10438+1.72317617j),(-23.16000+0.72512251j)
(-353.10438+1.72317617j),(-23.16000+0.72512251j)
Then I ran a code similar to yours, with a result similar to the following.
>>> np.genfromtxt('fname.txt', dtype=complex, delimiter=',')
array([[-353.10438+1.72317617j, -23.16000+0.72512251j],
[-353.10438+1.72317617j, -23.16000+0.72512251j],
[-353.10438+1.72317617j, -23.16000+0.72512251j],
[-353.10438+1.72317617j, -23.16000+0.72512251j]])
, , , .
!