Confounder adjusting before applying the ITE model to observational data

[YeechingTiger]

When we are using observational data for trial emulation and estimating the ATE, we use techniques like PSM or IPTW on the observational data to mimic the randomization in a randomized clinical trial for confounder adjusting.

When we are doing ITE, do we need to conduct a similar effort to get a matched/IPTW-adjusted dataset before we fit the data into the ITE model (e.g., meta learner)?