Inter-cell interference (ICI) in Multi-Cell Multi-User TDD systems,resulted from unavoidable reuse of pilot sequences across cells,bounds the performance of large-scale antenna systems (LSAS) in the infinite antenna regime. A new concept, called pilot contamination precoding (PCP), was recently proposed to cancel ICI. The main idea of PCP is that each base station linearly combines messages aimed to terminals from different cells that re-use the same pilot sequence. Crucially, the combining coefficients depend only on the slowfading coefficients between the terminals and the base stations. PCP allows building Multi-Cell LSAS in which SINR grows linearly with the number of base station antennas. In this work, we consider PCP design in the finite antenna regime,where three other types of interference become prominent. A balance among four types of interference is required. We formulate the PCP design as maximizing the minimum signal to interference-plus-noise ratio (SINR) subject to network sum power constraint. We further propose several algorithms for solving this max-min problem. Our simulation results show that in Linear Wyner-type cellular model with M=100 base station antennas the proposed algorithm give very significant, about 400 times, improvement in data transmission rate for 95% of users. Host: Lin Zhong