At Skima.ai, I upgraded a real-time AI interview platform built with Go, Python, and React.js, and optimized vLLM serving to reduce monthly cloud infrastructure costs by ~80%. I also migrated vector search to Turbopuffer and streamlined CI/CD and Kubernetes deployment workflows.
At Anvex.ai, I led an 18-member cross-functional team to deliver an enterprise Vision AI platform MVP in 20 days. I owned its MVP scope and backend architecture, including microservices, APIs, and real-time data pipelines.
At EaseMyAi Pvt. Ltd., I architected backend systems processing real-time data from 500+ cameras and IoT devices. I reduced API latency by 40% and MTTR by 50%, and delivered enterprise integrations for Indus Towers and DP World (Korea).
