Thanks for the FastAPI explanation. This makes sense. Right now, we use ELB (Elastic Load Balancer) to sit in front of multiple GPU instances. Is this suffici

Thanks for the FastAPI explanation. This makes sense. Right now, we use ELB (Elastic Load Balancer) to sit in front of multiple GPU instances. Is this sufficient or do you suggest adding Celery to this architecture?