If your MVP succeeds, you should probably throw away the majority of your architecture anyway, so you'll have the debt to address in the future regardless. Why not get up and running quickly?