Aayush Kumar Shandilya

About Aayush Kumar Shandilya

As a Software Engineer at Microsoft, I contribute to advancing service dashboards across…

Experience

1. Software Engineer

   Microsoft

   Aug 2023 — Present · Hyderabad, IN

   Contributed on service dashboards parity across Public and Sovereign (Air-Gapped) clouds, ensuring consistent observability across clouds• Optimized Data Egress Workflows: Drastically reduced cross cloud data egress on-boarding from 18 months to 4 weeks by re-engineering parity reporting via leveraging pre-approved schema to egress only parity deltas, successfully bypassing extensive sovereign cloud bureaucratic bottlenecks along with enabling other teams to fast track the same.• Engineered Automated SSL Rotation: Designed and built a polling-based SSL certificate rotation service from scratch. This service updates SSL certificates at runtime, enabling seamless renewals without server restarts or manual deployments on certificate expiration, maintaining high availability for API services.• Dashboard Parity Platform: Owned the end-to-end frontend implementation using React, TypeScript, and React Query. Integrated state management, advanced filtering/search, and Feature Flagging to enable safe, incremental production rollouts.• Architected Backend Reverse Proxy: Streamlined service integration by implementing a reverse proxy using YARP in the shared platform backend, eliminating the repetitive task of building custom controllers for every new route added in our service’s API.• Infrastructure & Security Migration: Spearheaded the transition from Azure Traffic Manager to Azure Front Door, enhancing global availability and security.• Hackathon Project:– LLM-Based Web Navigation: Developed agentic workflows using LLMs to predict and automate complex web browsing tasks (clicks, scrolls, form-filling).Hackathon Project: – Autonomous Dashboard Recommendation: Engineered a system utilizing LLMs to analyze service telemetry and automatically suggest missing observability dashboards, removing manual configuration friction.– LLM Spatial Awareness: Explored LLM architectures for vector-based, stroke-by-stroke image generation to evaluate AI spatial reasoning capabilities