This paper presents a novel cloud-integrated deep learning framework for the real-time core temperature estimation of automotive lithium-ion batteries (LIBs). Accurate core temperature estimation is crucial for rapid thermal management, thermal runaway prediction, and improving battery life and efficiency. The proposed framework integrates physical sensors, a cloud-based data acquisition system, and a Bidirectional Long Short-Term Memory (Bi-LSTM) neural network deployed on both the cloud and a local machine for real-time data collection, core temperature estimation, and visualization. This solution addresses the limitations of traditional temperature monitoring methods, enabling real-time processing and predictive capabilities. The paper discusses the model's architecture, experimental validation, and deployment strategies. The Bi-LSTM network achieves a core temperature estimation accuracy of 0.16°C, even with unknown cell data at varied ambient temperatures and Crates. Unlike existing state-of-the-art studies, this work provides the first demonstration of real-time core temperature estimation with an estimation accuracy of 0.31°C alongside the measurement and visualization for an entire battery module, marking a significant advancement in the field. Additionally, the paper demonstrates real-time decision-making informed by core temperature to prevent overheating and thermal runaway. This is achieved by implementing an automotive-grade CAN communication-based feedback control loop for the charging and discharging of the LIB module. The study also emphasizes the benefits of considering core temperature in improving the response time of thermal management by at least 2 minutes compared to surface temperature-informed control, which is extremely crucial for preventing overheating and protection from thermal runaway. A detailed discussion of computational costs and latency is included, providing a practical reference for realworld applications.