A longstanding goal in reinforcement learning is to build intelligent agents\nthat show fast learning and a flexible transfer of skills akin to humans and\nanimals. This paper investigates the integration of two frameworks for tackling\nthose goals: episodic control and successor features. Episodic control is a\ncognitively inspired approach relying on episodic memory, an instance-based\nmemory model of an agent's experiences. Meanwhile, successor features and\ngeneralized policy improvement (SF&GPI) is a meta and transfer learning\nframework allowing to learn policies for tasks that can be efficiently reused\nfor later tasks which have a different reward function. Individually, these two\ntechniques have shown impressive results in vastly improving sample efficiency\nand the elegant reuse of previously learned policies. Thus, we outline a\ncombination of both approaches in a single reinforcement learning framework and\nempirically illustrate its benefits.\n