Boiling heat transfer offers exceptionally high heat removal capability, making it attractive for energy-efficient thermal management in electronics cooling, power systems, and other high heat-flux applications. However, the practical use of boiling is constrained by the critical heat flux (CHF), beyond which vapor blanketing can trigger a sharp surface-temperature excursion and potential device failure. This talk presents an acoustics-based sensing and control framework for predicting and mitigating CHF in pool boiling systems. Acoustic emissions generated during boiling are used as a non-intrusive diagnostic signal to identify boiling regimes and detect precursors to CHF. Deep learning models trained on these acoustic signatures enable real-time classification of boiling states and advance prediction of impending CHF. Beyond prediction, the framework is extended to adaptive control, where an on-demand cooling intervention is activated based on acoustic feedback to delay thermal runaway and push the operating limit beyond the nominal CHF condition. The results demonstrate how low-cost acoustic sensing, machine learning, and active control can transform boiling from a passively monitored heat-transfer process into an actively regulated thermal management strategy. The broader implication is a pathway toward safer, more compact, and more energy-efficient two-phase cooling systems that operate closer to their physical limits.