Piyush Dubey — Principal Platform Software Engineer

visitor@piyush-dubey.is-a.dev: ~/portfolio — zsh — 120×40

● ssh

● portfolio — zsh

● about.md

● experience.log

● skills.json

● projects/

● contact.vcf

visitor@piyush-dubey.is-a.dev:~$

body{font-family:monospace;background:#0a0e14;color:#c9d1d9;padding:2rem;line-height:1.7}Piyush DubeyPrincipal Platform Software Engineer — Developer Infrastructure & Distributed Systems
Bangalore, India · Open to relocation & visa sponsorship · 13+ yearsAboutPlatform engineer with 13+ years setting technical direction for distributed systems and AI infrastructure at Oracle, Microsoft, and Adobe. Architected a retrieval platform serving 220M+ monthly users, owned an MCP-based AI agent execution platform cutting incident resolution time 30%, and turned per-team solutions into shared infrastructure adopted across 20+ engineering teams.ExperienceOracle — Principal Platform Software Engineer, Developer Infrastructure (Jan 2025–Present)

Built self-service Terraform Provider Registry collapsing release cycles from 3 months to 1 week (12×) across 20+ provider teams

Architected event-streaming platform: 5K+ events/sec, 430M+ events/day, <500ms latency, adopted across 20+ OCI teams

Designed Redis-backed distributed rate limiter: 10K+ QPS, <5ms p99, quota enforcement and tenant isolation

Shipped autonomous incident-resolution agent (Codex SDK + MCP): 60% less debugging, 30% faster resolution, 15-min recovery

Mentored 40+ engineers; technical owner of 6+ critical cloud services

Microsoft — Senior Software Engineer, Outlook Groups Platform (May 2022–Dec 2024)

Owned architecture of Outlook Groups retrieval platform powering Copilot search across Microsoft 365 (220M+ monthly active users)

Cut global retrieval and collaboration latency 70% via distributed caching and query-path redesign

Reduced enterprise infrastructure costs ~10% with lifecycle management capabilities

Adobe — Computer Scientist, Globalization Platform (Oct 2016–Apr 2022)

Designed centralized real-time globalization platform: 31 locales, 50K+ QPS, saving ~2 weeks of localization effort per release across 5+ product lines

Architected Kafka-based distributed event pipelines for reliable async processing

TCS — Software Engineer, Healthcare Data Pipelines (2013–2016)SkillsJava · Go · Python · TypeScript · Spring Boot · Micronaut · Dropwizard · FastAPI · Distributed Systems · Event-Driven Architecture · Workflow Orchestration · High Availability · OCI · AWS · Azure · Kubernetes · Docker · Terraform · Kafka · Redis · PostgreSQL · Cassandra · MongoDB · Qdrant · MCP · AI Agents · RAG · Vector Embeddings · OpenAI SDK · Codex SDKContactEmail: get.piyush007@gmail.com · LinkedIn: linkedin.com/in/pidubey007 · GitHub: github.com/GhOsT00712