Under ideal circumstances, is it possible to construct a 100% efficient photoelectric circuit? Lets assume some ideal circumstances:
1) The incident light has same frequency, greater than the
    threshold frequency (sufficient to eject inner electrons too), throughout. 
2) The work function of the metal
        doesn't change over time.
Is this setup able to eject an electron for each photon? or the electrons lose their way inside the atom and hence cannot escape?
 A: No it is not. There are other processes that can occur when a photon interacts with an atom. The photoelectric effect is governed by the photoelectric cross section. The image below shows how this cross section varies with energy and also shows the competing reactions that can occur.

Ref. http://rcwww.kek.jp/research/shield/photon_r.pdf
At some photon energies, the photoelectric effect may be the dominant reaction but the cross section is still not large enough to always eject an electron. The chart below shows the photoelectric cross section for different energies.

Ref:https://www.physics.queensu.ca/~phys352/lect17.pdf
A: The conditions, (1) and (2), that you've listed are not so hard to meet, but the efficiency of the photoelectric effect is not limited just by these conditions.
In other words, even if the frequency of the incident light is sufficient (1) and the work function is constant (2), and even if all photons are absorbed and pass their energy (exceeding the work function) to electrons, not all of these energized electrons will be emitted, since, as explained in this post, they may not necessarily end up moving in the direction of the exposed surface.
