The game of TIC-TAC-TOE can be modelled as a non-deterministic Markov decision process (MDP) if, and only if:
The opponent is considered part of the environment. This is a reasonable approach when the goal is to solve playing against a specific opponent.
The opponent is using a stochastic policy. Stochastic policies are a generalisation that include ...
Deep reinforcement learning might be suitable for your requirement. If possible I would recommend you to add DRL hashtag to your thread, so that it can reach to larger number of DRL researchers and practitioners, to help you out.