shiva raghav

I’m an early-career software engineer pursuing an M.S. in Computer Science, and I enjoy owning backend-heavy systems end-to-end—from design and implementation to production reliability. I’ve shipped production LLM and SaaS features in ambiguous environments by making pragmatic technical decisions and focusing on measurable outcomes.

At Mermory, I owned a production LLM ingestion pipeline for 500+ daily users and improved reliability from 70% to 98% using retries, idempotency guards, and fault isolation. I also built an OCR processing pipeline with Node.js child processes to reduce main-thread blocking and improve API response times by ~40%, and I shipped an autonomous document ingestion agent that processed 12,000+ items, eliminating ~80% of manual content creation work.

At Agnikul Cosmos, I decomposed a monolithic subsystem validation workflow into async GraphQL microservices, improving mission-critical throughput by 35%, and I built internal React + TypeScript tooling to reduce pre-launch triage time by 25%. I’ve also increased automated test coverage from 30% to 80% via unit, integration, and end-to-end tests in CI, and I’m currently running real-user projects like FlowTask, with 99.5% uptime backed by structured logging, queue health checks, alerting, feature flags, and CI/CD releases.