Priya Ramanathan

Priya spent four years at Vantage building cost-allocation tooling for AWS customers, then two years at HashiCorp on the FinOps side of Terraform Cloud billing. Before that she was a backend engineer at Twilio, where she rewrote the internal usage-metering pipeline that powered SMS billing for roughly 1.4 billion messages a day. She holds AWS Solutions Architect Professional and the FinOps Certified Practitioner credentials, and contributes irregularly to the OpenCost project. Her current obsession is Savings Plans portfolio math: she keeps a spreadsheet at home that models the break-even point between 1-year no-upfront and 3-year all-upfront commitments across 11 EC2 families. Priya writes mostly about EC2 rightsizing, S3 storage-class transitions, and the unglamorous work of tagging governance. She lives in Austin and is slowly losing a fight with her landlord over a second monitor stand.

Статті автора Priya Ramanathan

Як оптимізувати витрати на GPU для AI/ML у хмарі: економія до 80% у 2026
Стратегії та найкращі практики

Як оптимізувати витрати на GPU для AI/ML у хмарі: економія до 80% у 2026

Повний посібник 2026 року з оптимізації витрат на GPU для тренування та інференсу AI/ML у AWS, Azure, GCP. Spot GPU, квантизація FP8/INT4, vLLM, Bedrock vs SageMaker — і реальні приклади економії до 80%.

Priya Ramanathan 7 хв читання