At GobbleCube.ai, I build and scale Python/FastAPI backend services for an AI revenue platform processing 100M+ row datasets and 1M+ SKUs. I redesigned the queueing and rate-limiting layer, cutting operation error rates by approximately 90%.
I deploy services on AWS EKS using Helm and ArgoCD, and built Temporal workflows for automated bid placement and campaign execution across Q-Commerce and E-Commerce platforms.
I architected GobzIQ, a multi-agent analytics system served over MCP, and built a ClickHouse MCP server for LLM agents. I also built an autonomous on-call agent integrated with ClickUp and Claude Code.
In personal projects, I built a Go distributed transaction coordinator using gRPC and Two-Phase Commit, and Kafkaesque, a durable C++20 event broker. I was also a finalist in the Nationwide Supermind Hackathon, where I led a team building Social Lens.

