At Arintra, I overhauled an accuracy-benchmarking pipeline, cutting a 65k-case batch runtime by 5× and compute cost by 41%. I also traced downstream latency to its cause and reduced it by 96%.
I designed Kafka-based workflows for platform services processing 20k+ charts a day. I also shipped a production RAG pipeline and an AI guardrail framework for context-aware search and safer LLM interactions.
At Arintra, I designed a centralized RBAC microservice adopted across core services and built audit tools for chart-coding accuracy. The reporting service I developed contributed to improvements in targeted error buckets and audit SLA adherence.
Previously, at Aidash Systems India Private Ltd, I redesigned service orchestration around asynchronous event-driven microservices and helped develop batch inference and RabbitMQ-based services. That work improved throughput, reduced client wait times and API timeouts, and achieved 99.9% message delivery reliability.

