Hey Git, ever pushed a data pipeline to the edge and had to decide whether to tweak the algorithm or just refactor the code for speed? I’ve been thinking about how to squeeze every ounce of performance out of it. What’s your take?