Estimate heterogeneous treatment effects with causal ML

You have logged, observational data from an always-on promotion and want to target users to maximize incremental revenue. Propose a causal ML approach to estimate heterogeneous treatment effects (HTE) and derive a targeting policy. Specify: (a) choose among T-/S-/X-/DR-learners and justify with bias–variance and ignorability considerations; (b) model choices and cross-fitting to reduce overfitting and nuisance bias; (c) handling positivity/overlap violations and extreme propensities; (d) evaluation metrics (policy value estimation, uplift/Qini, PEHE proxy) and how to implement off-policy evaluation with IPW and doubly robust estimators; (e) fairness constraints (e.g., demographic parity of treatment) and their impact on policy; (f) how you would A/B test the learned policy safely before full deployment.