We present the Learned Ranking Function (LRF), a system that takes short-term user-item behavior predictions as input and outputs a slate of recommendations directly optimizing for long-term user satisfaction. Most previous work is based on optimizing hyper-parameters of a heuristic function. We propose to model the problem directly as a slate optimization problem with the objective of maximizing long-term user satisfaction. We also develop a new constraint optimization method that is able to improve the stability for multi-objective optimization. We evaluate our approach with live experiments and describe its deployment on YouTube.