Lead DevOps Engineer @ Simbian
I architect and operate cloud-native platforms — Kubernetes, Kafka, and observability at production scale, across every major cloud and on-prem.
Cloud-native engineer, platform architect, systems thinker.
I lead DevOps at Simbian, where AI agents take on cybersecurity — and where I own everything from multi-cluster, multi-cloud Kubernetes to shipping our product into customer on-prem and airgapped environments. Before that, I spent four years at Falkonry running multi-region infrastructure for time-series AI: Kubernetes, Kafka, Pulsar, HDFS, high-speed datastreams and a large open-source data stack under real industrial load.
My work sits at the intersection of platform engineering and architecture: building observability that scales, release processes that let one team ship 50+ microservices to many environments, and developer platforms that replace "works on my machine" with ephemeral, reproducible environments. Today that means keeping multiple Kubernetes clusters across three clouds above 99.5% uptime.
Outside of work I organize CNCF Mumbai tech events, am a Grafana Champion, speak at conferences and meetups, and contribute to open source in this way. I care about operational maturity — systems that are boring at 3 AM — and about getting things done across team boundaries.
$ status --current
6+
years in production
4
clouds operated
35+
Kubernetes clusters
100+
services on CI/CD
>99.5%
uptime
90%
MTTD improvement
The stack I run in production — not a logo wall, but tools I've operated, broken, and fixed at scale.
Production systems I've architected and operated — Kubernetes platforms, Kafka pipelines, observability stacks and developer tooling.
End-to-end observability across multi-cloud Kubernetes fleets — metrics, logs, traces and alerting built from scratch and operated in production.
Architecture
Per-cluster Prometheus in agent mode remote-writes to a central hub; Thanos adds HA, deduplication and long-term object storage; Grafana handles dashboards, alerting and cloud-provider plugins.
MTTD improved by 90%, MTTR by 50%
Operated and scaled Kafka and Pulsar clusters powering time-series AI pipelines — high-throughput ingest with strict ordering and retention guarantees.
Architecture
Multi-broker clusters on Kubernetes with rack-aware replica placement, workload-isolated topics, tiered retention, and consumer-lag SLOs wired into Prometheus alerting.
CI/CD and release automation for 50+ microservices deploying to Kubernetes clusters and cloud services across staging, testing and many production environments.
Architecture
Trigger-based release pipelines push artifacts and images to internal and customer registries; a Helm-based operator (Operator SDK) drives rollouts with pause/resume for database migration failures.
Scaled from one hand-rolled environment to many, fully automated
Feature-branch deployments and ephemeral test environments so developers ship and validate without touching shared infrastructure or local hacks.
Architecture
Automated environment lifecycle — creation, dependency wiring, database initialization, cleanup and reporting — for every application component, per branch.
Replaced local-only developer setups entirely
Automated container and OS-level vulnerability management feeding SOC2 compliance and customer-facing security reporting.
Architecture
Trivy, AWS ECR scanning and Clair (Quay) cover application images; OS package scanning with automatic fixes; version-diff analysis with automated report generation and delivery.
Configuration management and drift-free operations for cloud instances and the monitoring stack itself.
Architecture
Playbooks automate EC2 patching, Grafana dashboard and alert sync, and OpenSearch monitor provisioning — observability config managed as code.
I speak about FOSS, Kubernetes, Cloud, Observability, Platform Engineering, or really any piece of software that interests me - Kafka, Opensearch among others — grounded in production war stories, not slideware.
GrafanaCON Local PunePune, India
From zero monitoring to a robust multi-cluster Kubernetes observability stack — architectures from single cluster to multi-cloud, with pros, cons and use cases. Hosted at the InfraCloud office.
OpenSearchCon India 2025Bengaluru, India
Running OpenSearch effectively on Kubernetes takes a well-structured approach — how to deploy, configure and operate it. Presented at the Sheraton Grand Hotel, Bengaluru.
DevOpsDays Atlanta 2025Atlanta, USA
Ignite talk on what an internal development platform is, why it matters and how to build one — delivered at the Historic Academy of Medicine and livestreamed on YouTube.
Notes from production — observability, streaming, Kubernetes and the occasional incident retrospective.
Infrastructure questions, speaking invitations, architecture reviews, or mentoring — my inbox is open.
$ topmate --book-session
Career guidance, DevOps interview prep, architecture reviews, or deep dives on Kubernetes, observability and Kafka — book time with me directly on Topmate.