Let . Then is a r.v. such that . The idea is to procede like for the proof of Markov inequality , including the fact that .

You have . Now find an upper bound for the first term using , and an upper bound for the second term (similar to Markov). Then you're almost there... I let you give it a try.