Furthering our efforts on open AI innovation, Pearl enables researchers and practitioners to develop production-ready reinforcement learning AI agents that are tailored to a variety of production and research scenarios. This includes environments with limited observability, situations with sparse feedback, and safety-critical applications.