Using slate_decomp_q_agent in production

[ashishlal]

I am new to RecSim and Reinforcement Learning. How can I use a slate_decomp_q_agent trained using RecSim in production or simply in the real world outside the simulation environment?