Headline

Site Reliability Engineer @ Okta | Multi-cloud Kubernetes (EKS, GKE, kops) | K8s, Terraform, ArgoCD, Datadog | Building AI agents for SRE’s | Open to relocate

About

I keep large Kubernetes fleets and Kafka platforms running where downtime is a security event, not just a revenue one. I'm an SRE at Okta, running EKS, GKE and kops across AWS and GCP — ~2,500 pods, ~300 nodes, 3 regions, 10+ production products. I migrated 150+ containerised workloads onto EKS with zero-downtime cutovers, taking availability from 99.9% to 99.95%. Most of my recent work is building AI into operations: ▪ Agentless AI incident triage — Designed the Kubernetes layer for an Amazon Bedrock/Claude on-call agent (MCP) that reads 50+ runbooks and correlates pod health, logs and events with Datadog and CloudTrail into one Slack root-cause summary. Resolution time down 25%, escalations down 20%. Built it agentless: read-only, auto-rotating ServiceAccount credentials and cross-account STS AssumeRole reach the whole kops fleet with zero VPC peering and nothing in-cluster to deploy, patch or be paged for; RBAC ships via ArgoCD ApplicationSets. ▪ SOC 2 evidence automation — Claude-driven framework turning a Jira ticket and control number into an audit-ready bundle: ~21 collectors across ~14 controls spanning AWS, GitHub, Jira, Datadog, PagerDuty and Kubernetes, YAML-mapped and git-tracked with UTC provenance. Days of screenshotting, down to minutes. ▪ Zero-trust supply chain — Moved image and Helm OCI pulls off Artifactory/GCR to AWS ECR across 20 clusters; 16 OIDC-federated IAM roles (IRSA) replaced static pull secrets, via tier-gated ArgoCD promotion, zero downtime. ▪ GitOps & IaC — ArgoCD + GitLab CI for 200+ microservices: deployment failures down 40%, release velocity up 30%. Terraform modules cut environment setup from 2 days to under 3 hours. ▪ Observability & SLOs — Datadog, Prometheus, CloudWatch and GCP Monitoring consolidated into one pane; 200+ alert rules tuned, false-positive pages down 35%, MTTD improved 30%. ▪ Kafka — A year at Confluent containerising brokers on Kubernetes with multi-AZ DR (RPO minutes to under 30s) and JVM/kernel tuning for +25% throughput. Confluent Certified Administrator + Developer. Best fit for regulated, high-availability infrastructure — identity, payments, streaming data, banking — where reliability, auditability and least privilege all have to hold at once. Open to Senior SRE / Platform Engineering roles in Singapore, Europe and India. Happy to relocate. mayankofficial2504@gmail.com Kubernetes · AWS · GCP · Terraform · ArgoCD · GitOps · Kafka · Python · Datadog · Prometheus · SLO/SLI · IRSA/OIDC · Zero trust · SOC 2 · Amazon Bedrock · MCP · AIOps

Location

Bengaluru

AI stack

claude code

claude

cursor

gemini

github copilot

openai

Work experience

Site Reliability Engineer at Leading Identity Provider MNC, Aug 2025 - Present · 1 yr 2 mos, Bengaluru, • EKS Migration at Scale: Migrated 150+ containerised workloads to AWS EKS across 3 regions (2,500+ pods, ~300 nodes) with zero-downtime cutovers, lifting availability from 99.9% to 99.95%. • Agentless AI Incident Triage: Designed the Kubernetes access layer for an Amazon Bedrock/Claude on-call agent (MCP) that reads 50+ runbooks and correlates pod health, logs and events with Datadog and CloudTrail into a root-cause summary — resolution time down 25%, escalations down 20%. Built agentless: read-only, auto-rotating ServiceAccount credentials and cross-account STS AssumeRole reach the kops fleet with zero VPC peering and nothing in-cluster to patch or be paged for; RBAC ships via one reviewed ArgoCD definition. • SOC 2 Evidence Automation: Turns a Jira ticket and control number into an audit-ready evidence bundle — ~21 collectors across ~14 controls spanning AWS, GitHub, Jira, Datadog, PagerDuty and Kubernetes, YAML-mapped and git-tracked with UTC provenance. Days of screenshotting down to minutes; audit findings down 30%. • Zero-Trust Image Supply Chain (ECR + IRSA): Moved image and Helm OCI pulls for 20 clusters across multiple AWS regions off Artifactory/GCR to AWS ECR; 16 OIDC-federated IAM roles replaced static pull secrets with short-lived STS tokens across 55+ deployments, via tier-gated ArgoCD promotion, zero downtime. • GitOps CI/CD: ArgoCD + GitLab CI for 200+ microservices; deployment failures down 40% and release velocity up 30% via automated rollbacks. • Multi-Flavour Kubernetes Ops: Operate EKS, GKE and kops clusters across AWS and GCP for 10+ products; zero-downtime version upgrades at full CIS-benchmark compliance. • Terraform IaC: Reusable modules for VPCs, IAM, EKS and GKE across 4+ accounts; environment setup from 2 days to under 3 hours, config drift eliminated. • Unified Observability & SLOs: Consolidated Datadog, Prometheus, CloudWatch and GCP Monitoring into one pane; 200+ alert rules tuned, false-positive pages down 35%, MTTD improved 30%.

DevOps Consultant 2 at Confluent, Sep 2024 - Aug 2025 · 1 yr, Bengaluru, - Led large-scale Kafka platform modernization by migrating critical components to Kubernetes, delivering automated scaling, high availability, and disaster recovery while reducing RPO to seconds. - Architected and optimized high-throughput, low-latency Kafka environments through deep systems tuning (CPU, memory, kernel), improving peak performance by 25%. - Drove platform automation using Python, Shell, Ansible, and Terraform, enabling repeatable migrations and cutting environment provisioning time by 50%. - Established end-to-end observability across Kubernetes and infrastructure layers using Prometheus, Grafana, and Splunk to improve reliability and operational insight. - Led complex incident response and root cause analysis across distributed systems, networking, and virtualization, consistently reducing MTTR in 24/7 production environments. - Partnered with engineering, SRE, and architecture teams to deliver scalable, resilient platforms aligned with SLAs and long-term business goals.

DevOps Consultant 2 at Cisco, Sep 2023 - Aug 2024 · 1 yr, Hybrid, - Network & Proxy Automation — Built Ansible playbooks and Terraform modules automating feature rollout to customer networking devices and managing DNS and L7 proxy configuration across global customer environments; cut deployment time 40% and eliminated manual configuration errors. - CI/CD Pipeline Engineering — Designed Jenkins pipelines with integrated test and validation gates for internal tooling adopted across Cisco's global engineering ecosystem; reduced rollback frequency 30% and P1/P2 incidents 20% over 18 months, instrumented end-to-end with Prometheus and CloudWatch. - Production Database Reliability — Owned MySQL and MongoDB across three production sites for CNAAP, tuning schema, indexing and query paths to hold performance through IPL-season peak traffic; 40% faster data processing with no customer-visible degradation. - AWS Platform Delivery — Deployed and operated the TOP/SLM service-level dashboard on AWS via Infrastructure as Code and CI/CD, replacing manual release steps and giving customer-facing teams live service health visibility.

DevOps Consultant at Cisco, Aug 2022 - Sep 2023 · 1 yr 2 mos, Bengaluru, - Automated the report generation pipeline for a customer-facing programme, removing a recurring manual step from the delivery team's weekly cycle. - Built the Automation Capture Portal, an internal tool tracking automation adoption across the department — used by leadership to identify where engineering toil was concentrated.

Technical Undergraduate Intern at Cisco, Jan 2022 - Aug 2022 · 8 mos, Bengaluru, - Contributed to Config Search Engine, an NLP-based tool mapping device configurations to their underlying features.

Education

Bachelor of Technology - BTech, Electrical, Electronics and Communications Engineering, LNCT Group of Colleges

High School, Physics,Chemistry,Mathematics, The Aditya Birla public school