Over the next decade, large scale structure (LSS) surveys will play an increasingly important role in the measurement of cosmological parameters and as a probe of initial conditions. Relating the observed galaxies to the physics of the early universe requires a good understanding of several late-time sources of non-linear effects. In this talk, I will explain how, looking at large scales, effective field theory (EFT) offers a natural framework to deal with these non-linearities. After reviewing the recently developed effective field theory of large scale structure, I will explain how the halo bias model - a model relating the density of dark matter halos to the underlying dark matter density - can be understood in the context of EFT. In the final part of the talk, I will show explicit computations of halo statistics and discuss their observational consequences.