Engineered a scalable text classification solution on AWS by training a custom model in Amazon SageMaker and deploying it through a production-ready serverless architecture for real-time inference."

Designed a production‑grade Kubernetes platform across two AWS regions with automated failover, GitOps deployments, and centralized observability.

Built a self‑service portal for developers to spin up ephemeral environments on demand using Terraform modules and GitHub Actions, reducing lead time for testing from days to hours.

Re‑architected a batch data pipeline to use serverless and spot instances, cutting monthly costs while improving throughput and reliability.

Defined and rolled out an observability baseline (logs, metrics, traces) for microservices, enabling faster incident response and data‑driven performance tuning.

Implemented a secure landing zone with guardrails, network segmentation, and identity standards, providing a compliant foundation for multiple product teams.