Software Engineer · Platform & DevOps
Gates Babu
I build cloud-native platforms and CI/CD systems that cut incident MTTR by 55% and deployment toil by 60% — and apply the same rigor to production-grade RAG and AI systems. IIIT Lucknow CS graduate and national hackathon winner.
Experience
Software Engineer — Platform & DevOps
Hashout, Bengaluru
March 2025 — Present
- Architected end-to-end CI/CD pipelines for 5 enterprise platforms using Jenkins Shared Libraries, Docker, and dynamic Kubernetes agents, increasing deployment frequency by 70% and eliminating manual handoffs.
- Engineered a differential build system in Python (via git-diff) to compile only changed microservices, reducing build latency by 20% and lowering daily compute costs without impacting delivery cadence.
- Standardized Kubernetes workloads via multi-stage Docker builds and immutable artifact promotion through JFrog Artifactory, cutting release toil by 60%; automated GCP Cloud SQL Proxy access to eliminate all hardcoded secrets.
- Built a ChatOps observability framework parsing CI/CD logs for root-cause signals and triggering real-time Slack alerts, reducing MTTR by 55% across production incidents.
Projects
UHC Policy AI Agent
Built a production-deployed AI agent using Hybrid RAG — ChromaDB vector search paired with Tavily live web retrieval — delivering clinically accurate, source-traceable answers over UHC insurance policy documents. Implemented a multi-tier fallback hierarchy (local vector search → live web retrieval via Gemini 2.5 Flash) for robust coverage outside the indexed corpus, with zero manual intervention.
Real-Time Multiplayer Tic-Tac-Toe
Developed a server-authoritative multiplayer game on Nakama with real-time WebSocket gameplay, private room codes, auto-matchmaking, and invite links — all game logic enforced server-side to prevent client-side cheating. Deployed Nakama + CockroachDB on DigitalOcean via Docker Compose with a Cloudflare Tunnel for HTTPS/WSS; shipped a React frontend on Vercel with GitHub CI/CD, a live leaderboard, and a 30s turn timer.
Patient Management System
Engineered a 4-service microservices architecture (Patient, Billing, Analytics, Auth) using gRPC for low-latency inter-service communication and Kafka for async event streaming, maintaining 99%+ availability at scale. Released on AWS ECS via CloudFormation IaC, secured every endpoint with JWT + API Gateway, and containerized all services with Docker — cutting deployment time by 30%.