At Brudite Private Limited, I lead backend development for a real-time train-tracking platform. I work with Kinesis Data Streams, Python AWS Lambda functions, and API Gateway WebSockets to deliver live movement updates.
I cut AWS event-processing costs by 30% by replacing SQS-based transport with Kinesis, and reduced historical movement query latency by over 97% through PostgreSQL indexing and table partitioning. I also moved data preparation from the browser into backend APIs and added validation and retry-aware processing.
I’ve built data pipelines and OpenSearch-powered search, and cut AI inference costs by moving to a local CPU-based runtime. On the deployment side, I migrated Docker workloads to Kubernetes and Helm, improved deployment times with CI/CD and Terraform, and built observability for serverless and containerised workloads.
