The mission of this project is to develop an ML-enabled system for tracking and positioning Low Earth Orbit (LEO) satellites that use uncooperative or undisclosed signal formats. Because LEO satellites are managed by private entities, their downlink specifications are often kept secret, which causes traditional physics-based orbit models to fail due to a lack of accurate initial estimates. To address this, the project reframes blind Doppler tracking as an applied machine learning task, using a Python-based system to estimate orbital parameters directly from delay time series data collected during partial overpasses. Additionally, the system aims to detect and classify anomalous events, such as orbital maneuvers, noise bursts, and data dropouts, to enhance the reliability of satellite monitoring and passive orbit validation.