Liquidity risk has emerged as one of the most critical threats to financial system stability, particularly in increasingly interconnected markets where shocks can propagate rapidly across institutions, asset classes, and geographies. Traditional liquidity measurement techniques often rely on static balance-sheet indicators or linear sensitivity models, which fail to capture nonlinear spillovers, dynamic feedback loops, and cross-market contagion. With growing market complexity and the rise of high-frequency trading, funding dependencies, and tightly coupled financial infrastructures, more advanced analytical tools are required to measure how liquidity stress travels through the system. Machine-driven quantitative modeling offers a powerful framework for detecting, quantifying, and forecasting liquidity risk propagation with far greater precision. This study examines machine-driven quantitative modeling approaches designed to measure liquidity transmission across interconnected financial systems. Broadly, the discussion begins by outlining the limitations of traditional liquidity metrics such as bid-ask spreads, coverage ratios, and static liquidity buffers in environments prone to rapid regime shifts and correlated withdrawals. It then shifts to advanced techniques capable of modeling systemic liquidity behaviour, including nonlinear network models, agent-based simulations, and machine learning algorithms that identify hidden propagation channels and emergent liquidity clusters. These methods allow for the identification of transmission pathways that may not be directly observable through conventional financial indicators. Narrowing the focus, the study explores how machine-driven models such as graph neural networks, stochastic process learning, and reinforcement-learning-based stress simulators capture dynamic liquidity interactions under stressed conditions. These tools enable more accurate forecasting of liquidity drains, intraday funding pressures, and cascading margin calls. By incorporating real-time market data, adaptive learning processes, and systemic interaction structures, these models provide early-warning insights that strengthen liquidity management frameworks. The paper concludes with practical considerations for implementation, data governance, and integrating machine-driven liquidity analytics into regulatory oversight.