Using Strategic Repartitioning, Caching & Checkpointing, Spark Tuning and Holistic Optimisations to improve spark job execution times by ~36% and improve cluster stability and efficiency without changing core business logic.