Foundations Of Real-World Reinforcement Learning