Worked on reliability improvements in the sapcc/nova repository by developing a periodic cleanup mechanism for virtual machines stuck in the DELETING state. Addressed the issue of lost delete requests by implementing a Python-based task that detects the absence of a process lock and automatically re-initiates the termination process. This backend development effort reduced the need for manual intervention, improved resource lifecycle management, and enhanced system reliability. Leveraged skills in distributed systems, cloud computing, and system administration to ensure that unrecoverable VM states are proactively resolved, contributing to more efficient and stable operations within the sapcc/nova cloud infrastructure.