Rudra Sikri

I’m a backend engineer with 2.5+ years shipping production Python and applied-LLM systems, including multi-agent orchestration on Microsoft autogen-core. I design runtimes and libraries that turn agent ideas into reliable, tenant-isolated services.

At CRUV Dimension Pvt. Ltd., I architected an AWS SQS-backed workflow runtime that routes business leads to AI agents across WhatsApp, voice/telephony, and email—driving 15%+ lead conversions across 10+ enterprise clients. I also consolidate services into a shared AI core library using a ports-and-adapters architecture, so teams can plug in their own LLM, storage, and outbound messaging.

I build scalable, correlated email-AI capabilities with FastAPI and per-account async IMAP IDLE, achieving 100% email-reply correlation accuracy via Gmail Plus Addressing. For orchestration, I delivered a 15-agent DAG-based runtime on autogen-core with checkpointing and MongoDB-backed two-level memory transfer patterns into production.

Earlier, I shipped backend end-to-end for Craver (Django/DRF + PostgreSQL/MongoDB/Elasticsearch), including auth, catalog, cart, order flows, grievance, payments (CCAvenue), and settlement. I’m motivated by async reliability, clean architecture, and measurable outcomes—conversion lifts and operational correctness.