We present a new two-phase augmented Lagrangian filter method that has the potential to exploit scalable parallel subproblem solvers. In the first phase, we approximately minimize the augmented Lagrangian to estimate the optimal active set. In the second phase, we solve an equality-constrained QP. An augmented Lagrangian filter determines the accuracy of the augmented Lagrangian minimization and ensures global convergence. We present a convergence analysis and preliminary numerical results.