Setting up a hierarchical model for paired t-test

Hi devs

I’m trying to do a paired t-test using Bayesian Inference for which I need to induce the following hierarchy

a[i, j] = a[j] + d_{ij}\;\;\;\; \forall i\in\{1,2,3,4\} and j\in \{1,2\}

where I want d_{ij} = 0 if j = 0 and d_{ij} \sim N(0,5) if j = 1

Basically, I want d to be learnt from the differences between two treatments j=0 and j=1.

so, a[0] is the baseline, and d measures the difference from baseline.

Could you please help me how I can write this into code with a (generic) plate notation?