I am not sure such a point K exists in the general case, but it may be possible to find such a point for this particular triangle. I have sketched triangle ABC here:
I think you should try to imagine this as an animation, moving each of the parallel line segments toward their intersection (indicated by arrows). I have drawn arrows, but keep in mind that all three segments will also intersect, say, to the right of the segment parallel to AB.
I'm not sure how to solve it yet, but I would probably start by finding the slopes of the lines parallel to AC and AB, solve the system for their intersection. Then, once you have that (there will be an infinite number of points of intersection within triangle ABC), you can use the line parallel to BC (slope = 0 in the diagram) to find the distance, since the
y values in all coordinate pairs on that line will be the same for every case of the line.
Looking at the picture, it seems the most likely candidate for a point of intersection where all three segments have the same length will be at a pretty low value of
y. If you think about it graphically, the premise of the problem stipulates that all three segments have the same length at some point of intersection. The two non-horizontal segments in my picture will only meet at certain points where they both have the same length. That sounds like you should set their distances equal to each other. If that is indeed how this scenario plays out, I sure hope that horizontal line segment is the same length at one of those points. If not, there can be no solution (using real numbers).
You might also find some interesting theorems about triangles and the lines within them at
http://www.pballew.net/medians.htm
Keep at it. Nice problem, if there's a solution.....