This independent case study explores how FinOps principles can be applied to reduce cloud costs in real-time AI inference workloads.
Using a simulated anime recommender system deployed with AWS SageMaker, Lambda, and related services, this memo analyzes unoptimized vs. optimized configurations through tools like AWS Cost Explorer, Pricing Calculator, and CloudWatch.
The result: a projected 90%+ cost reduction, achieved through architectural changes, instance right-sizing, and usage monitoring.
🔒 Independent project. Not affiliated with any employer.