Using high frequency individual-level panel data from India, we show that in- come inequality, measured as the ratio of high-skilled to low-skilled income, in- creased sharply following the imposition of lockdown triggered by COVID-19. To explain this fact, we integrate a canonical SIR epidemiological model into a general equilibrium framework with high-skilled and low-skilled workers, each choosing to work either from their work locations (onsite) or from their homes (remote). On- site and remote labour are imperfect substitutes, but more substitutable for high- skilled relative to low-skilled workers. Upon introducing pandemic containment policies such as lockdowns that restrict labour mobility, the model can explain around 60 percent of the observed increase in inequality. We also show that the containment policies are less effective in controlling disease spread among low- skilled workers as they optimally choose to work more onsite compared to their high-skilled counterparts. Introducing direct transfers for low-skilled workers re- verses this increase in inequality and reduces the disparity in health outcomes be- tween high-skilled and low-skilled workers.