The rapid growth in computational demands driven by advancements in artificial intelligence, machine learning, and quantum computing are leading to significant surge in energy consumption in data centers (DCs). Recent reports indicate a 31% increase in DC energy usage in 2022, with an overall 400% rise since 2015, primarily attributed to the deployment of AI-driven workloads and high server redundancy configurations (e.g., 2N + 1) to achieve 99.99% uptime in Tier-3/4 facilities. This redundancy maintains continuous power requirements, even when a substantial portion of CPU cores are idle. To manage thermal stress under full load, advanced cooling methods, including liquid cooling and large-scale air-cooled systems, are implemented, often accounting for over 50% of total operational costs. This paper introduces a software-centric solution for granular thermal management at the individual CPU core level, incorporating temperature-aware job scheduling and placement strategies. This paper introduces a hybrid thermal-aware scheduling policy for HPC systems. Results demonstrate a 30% reduction in instantaneous core temperatures and 50% instantaneous power savings, optimizing energy efficiency while preserving performance. A novel hybrid throttling policy selectively manages thermally active jobs (hot jobs), effectively delaying the need for conventional DTM and DVFS interventions while ensuring efficient energy utilization for other jobs in the execution queues.