Your question is Fine-Tune and Serve Firefly Models. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How would you design a production system for fine-tuning and serving custom enterprise Adobe Firefly generative models under strict latency constraints?