This workshop will focus on opportunities for incremental optimizations using GPUs in scientific computing workflows. This is targeting instances where researchers have legacy codes, want to adopt existing frameworks, or are looking to integrate GPU compute into their own workloads. We will focus on four main categories of optimization. This session will highlight practical steps for CEOAS researchers to incorporate GPU acceleration incrementally, enabling impactful performance gains while working within existing toolchains. Prerequisites: No advanced programming experience required, however we will be diving into coding examples during the workshop. Attendees are encouraged to bring a laptop.