What in the world are you working on that outputs a 1nS -1kV pulse?
I hope you realize that a 1nS pulse will require a circuit with a bandwidth >= 10GHz in order to make a reasonable facsimile of the pulse, but inverted, right? That's in the radar X-band RF spectrum. At those frequencies, a tiny kink in a wire acts like an inductor. It's even difficult to use multilayer ceramic caps due to their parasitic inductance.
In the beginning of your thread, you posted a waveform that's practically identical to the one in this thread: http://forum.allaboutcircuits.com/showthread.php?t=59546
and the OP is wanting to invert the signal, stretch the pulse, and detect it using a microcontroller.
The pulser i am working on is not the one as shown in the beginning of the thread. But the structure of the output stage is the same-- a capacitor and a load resistor. The current flows from GND to the load. So that pulse is negative.