Sabya Sachi

Describes evolving an energy forecasting stack from SARIMAX/seasonal decomposition to machine-learning and deep-learning approaches; architecture includes TimescaleDB, FastAPI, MLflow, and approaches for scaling and continuous improvement.

Explains the migration from an SQS+Flask replication pipeline to AWS DMS using CDC for near-real-time MySQL→Postgres replication, lessons learned about monitoring, schema evolution, and cost trade-offs.

A guide sharing experience and advice for preparing for the AWS Solutions Architect certification.