Shriya Nayyar

At AgentStraw, I'm architecting an end-to-end compliance automation platform on GCP that replaces a manual government-portal workflow with a one-click pipeline. I built an async job engine that processes 90+ batches per daily run while preventing double-processing through Firestore checkpointing, normalization, and idempotent retries.

I also deployed a five-service GCP infrastructure on Cloud Run, VPC, Cloud Router, Cloud NAT, Secret Manager, and a reserved static IP to support third-party API allow-listing without interruption.

Previously at TinyWires Technologies, I delivered five-plus full-stack web applications on GCP and cut delivery cycles by 30%, owning architecture, backend APIs, and React/JavaScript implementation.

Outside internships, I build systems-heavy and AI products including MiniServe, a continuous-batching LLM inference server; WorkQueue, a Redis-backed distributed job system; and MoodMuse, an NLP-driven music recommendation engine serving personalized recommendations across 500+ songs.