Automated Planning is one of the founding research areas of Artificial Intelligence and Temporal Planning is the problem of finding a plan in a system where timing is not negligible and subject to temporal constraints. In this PhD research program, collocated in the context of the STEP-RL ERC project, the candidate will research novel Reinforcement Learning and Planning approaches for specializing different temporal planning algorithms by learning heuristics and other forms of search bias and by adapting the algorithm to retain formal and/or statistical guarantees. In addition to state-space search algorithms, we will adapt other types of planning algorithms such as symbolic encodings and plan-space search.