> whoami

Piyush Dubey

I build reliable platforms, high-performance APIs, and distributed systems that operate across multiple regions. Most recently, that means shaping cross-region delivery infrastructure at Oracle OCI and practical AI tooling that helps engineering teams move faster.

./about
./selected_impact
./notes
connect_linkedin
git_clone_repo
01.

About

> Loading current focus...

I am a Bangalore-based engineer working across platform engineering, backend systems, and applied AI. I like building the things that sit on the critical path: release orchestration, event pipelines, search systems, internal tooling, and the APIs that tie them together.

Right now, I work as a Principal Member of Technical Staff at Oracle OCI, shaping developer experience and continuous delivery infrastructure at global scale. The work spans orchestration services, cross-region architecture, Kafka-backed event flows, reliability improvements, and latency reduction where it has visible impact.

Across Oracle, Microsoft, Adobe, and earlier enterprise systems work, the pattern has stayed consistent: take an operationally messy problem, design a system that can scale, and make it simpler for both users and engineers. Lately that also includes AI agents, MCP-based tooling, and practical automation that reduces turnaround time instead of just adding hype._

./capabilities --list

  • Java
  • Python
  • Go
  • C#
  • Kafka
  • Docker
  • Kubernetes
  • PostgreSQL
  • Oracle
  • Cassandra
  • Redis
  • LangChain
  • MCP
  • OpenAI GPT
  • Codex
  • Claude
  • OCI
  • AWS
  • Azure
  • REST
  • Microservices
  • Event-Driven Systems
  • System Design
02.

Selected_Impact

A few areas where I have spent most of my time recently: platform reliability, developer productivity, and backend systems operating at real scale.

Cross-Region Release Platform

Oracle OCI | Platform Engineering

  • I build orchestration services and APIs behind OCI's continuous delivery platform, supporting deployments across 25 regions and roughly 100K monthly releases.
  • I designed Kafka-backed event flows for CI/CD workloads so the platform could support cross-region delivery with strong durability guarantees.
  • I also pushed reliability and latency work, including service decomposition and targeted caching for high-traffic cross-region API paths.

AI Tooling That Engineers Actually Use

Oracle OCI | Applied AI + Developer Experience

  • I built internal AI agent workflows for operational support and developer tooling using LangChain, MCP, OpenAI GPT, Codex, and Claude.
  • The goal was faster execution, not novelty, and the result was materially lower turnaround time and less engineering effort for the team.
  • That work also included Python and Java tooling, MCP server support, and mentoring engineers on where AI helps and where it just adds noise.

Search and Collaboration at Microsoft Scale

Microsoft 365 | Backend + AI Platform

  • I owned architecture for the Groups AI Platform powering semantic search and Q&A experiences for more than 220M monthly active users.
  • I drove API and retrieval optimizations that improved group creation latency by 70% and led execution across Outlook, Teams, and SharePoint.
  • I also shipped lifecycle capabilities that helped enterprise customers reduce subscription cost overhead.

Globalization, Messaging, and Data Foundations

Adobe + Earlier Systems Work

  • At Adobe, I built real-time translation services supporting 31 locales at 50K QPS, along with Kafka-based event pipelines designed for retries and dead-letter handling.
  • I also worked on globalization and messaging integrations that helped teams ship internationally faster and drove strong adoption in the Japan market.
  • Earlier in my career, I built healthcare data workflows and learned the foundations that still shape how I approach reliability, validation, and delivery.
03.

Notes

I write occasionally about platform engineering, distributed systems, backend architecture, and practical AI for developer productivity.

read_on_linkedin