Atul Singh

DevOps / SRE Engineer with 5+ years running production Kubernetes and Docker platforms on AWS, GCP, and Azure; provisioning infrastructure with Terraform; and managing Amazon RDS PostgreSQL and MongoDB at scale. Defines SLOs, SLIs, and error budgets to sustain 99.95% uptime, leads incident response with structured runbooks, and ships MLOps pipelines (LLM + Qdrant RAG). Serves as a technical point of contact for customers and stakeholders, iterating rapidly on integrations and delivering 40% faster deployments, 75% faster threat detection, and SOC 2 compliance.

Immediate joiner, open to remote, hybrid, and relocation opportunities.

DevOps Engineer II — Infrastructure & Security Team - KlearNow - Gurgaon, India

(2025-09)

AI Document-Extraction & Platform Reliability

• Operated production Kubernetes + Docker workloads on ArgoCD GitOps platform with SLO/SLI tracking, automated rollback, drift detection, and progressive delivery — sustaining 99.95% uptime and improving release reliability by 40%.
• Modernized CI/CD from Jenkins to GitHub Actions with reusable workflows, build serialization, and dependency-aware ordering — cutting deployment time by 30% and failed/conflicting deployments by 90%.
• Led zero-downtime major-version upgrades of Amazon RDS PostgreSQL and MongoDB across production — validated schema/driver compatibility with engineering teams, cutting upgrade downtime to near zero.
• Designed and shipped an MLOps pipeline for AI-driven logistics document extraction using LLM APIs (OpenRouter/Claude) and Qdrant for vector-based retrieval — reaching 80% production accuracy with scalable inference workflows.
• Served as technical point of contact for customers and stakeholders on the AI pipeline — gathering requirements, iterating on prompts, retrieval logic, and integrations to meet accuracy and turnaround targets (Forward Deployment model).
• Built LLM observability platform (Datadog, OpenTelemetry, SigNoz) for distributed tracing, APM, and token-level analytics — cutting inference issue detection time by 70%.
• Architected cloud security automation using AWS GuardDuty + EventBridge + Lambda + SNS, cutting mean time to detect multi-region threats by 75%.
• Supported SOC 2 compliance (OneLeet, TruffleHog); integrated Microsoft Entra ID with AWS ALB via OIDC for centralized Kubernetes ingress auth; implemented DR runbooks with RTO/RPO targets and automated quarterly failover testing.
• Evaluated Karpenter for EKS node autoscaling — replacing static EC2 node groups with Pod-requirement-aware provisioning, Spot Instance bin-packing, and automatic de-provisioning of underutilized nodes — targeting 25–30% reduction in cluster compute costs.
• Managed Helm chart releases across dev, staging, and production — maintained chart versioning, per-environment values overrides, and ArgoCD ApplicationSet configurations for consistent multi-env GitOps deployments.
• Wrote Python scripts and tooling to automate operational workflows and support microservice integrations for the AI document-extraction pipeline, reducing manual intervention across repeatable tasks.

SDE I / DevOps Engineer - Jio Platforms Limited - Gurgaon, India

(2023-01 - 2025-09)

Enterprise GitOps & Platform Security

• Migrated 150+ CI/CD pipelines (Jenkins → GitHub Actions, OIDC auth) — cutting deployment time by 40%; authored reusable Terraform modules for multi-environment provisioning, cutting new-environment setup time by 80%.
• Defined and monitored SLOs/SLIs; drove toil reduction via runbook automation and self-healing CronJobs — cutting manual operational work by ~30% and MTTD by 35% through Prometheus/Grafana dashboards.
• Launched enterprise GitOps platform on ArgoCD (App-of-Apps) achieving 99.95% system reliability; integrated OPA policy-as-code + Trivy into CI/CD and hardened clusters via kube-bench (CIS Benchmark) for SOC 2 readiness.
• Led in-place and blue/green major-version upgrades of Amazon RDS PostgreSQL and MongoDB replica sets with automated backup/restore and failover testing — ensuring zero data loss.
• Engineered Golang and Java microservices handling 1M+ daily transactions, writing core service code and improving system latency by 35%.
• Rolled out Kafka-based event-driven architecture (TLS/SASL) integrated with PostgreSQL and MongoDB — improving scalability by 50% and real-time processing throughput by 35%.
• Built distributed LLM observability (Datadog, OpenTelemetry, SigNoz), cutting incident diagnosis time by 40%; automated server config via Ansible playbooks for repeatable, consistent Linux environment builds.
• Analyzed AWS Cost Explorer data and implemented right-sizing and Reserved Instance strategies — reducing cloud spend through Spot Instance adoption, Savings Plans, and idle resource cleanup.
• Administered Bitbucket with trunk-based branching; enforced code quality via SonarQube static analysis and Fortify security scans — maintaining compliance and security standards across all repositories.
• Led capacity planning and load testing using k6 — identified bottlenecks in API gateways and database connection pools, informing autoscaling policies and Reserved Instance purchase decisions.
• Deployed and managed Kong API Gateway for centralized traffic routing, rate limiting, and authentication across microservices; integrated Kuma service mesh for mTLS, traffic policies, and inter-service observability.
• Wrote Python scripts for automation and microservice tooling — including operational scripts for deployment workflows and data synchronization tasks — alongside Golang/Java service development.

Software Developer - Sterlite Technologies Ltd - Gurgaon, India

(2021-12 - 2022-12)

Microservices Platform & Monitoring · Acquired by Jio Platforms, Jan 2023

• Shipped Java and Golang microservices across 20+ Kubernetes clusters on GCP using Helm and CI/CD pipelines; wrote Golang microservice code using goroutines and channels, improving throughput by 30%.
• Optimized PostgreSQL performance by 25% through query tuning, indexing, and connection-pooling; implemented Prometheus + Grafana monitoring stack reducing MTTR by 30%.
• Automated Linux server provisioning using Chef recipes and Bash scripts — standardizing environment configuration across development, staging, and production for consistent, repeatable deployments.

Machine Learning Engineer — Intern - Windhya Cloud Works Services Pvt Ltd - Remote

(2021-03 - 2021-05)

COVID-19 Analysis using Quantitative Model Approach

• Built a COVID-19 trend analysis project in Python using a quantitative modeling approach, applying statistical and machine learning techniques to identify patterns in pandemic data.
• Combined ML and DevOps practices to operationalize the project — writing Kubernetes manifests to containerize and deploy the model pipeline on GCP for reproducible, repeatable execution.
• Gained early hands-on exposure to core DevOps fundamentals — basic CI/CD automation, Docker containerization, and version-controlled deployment workflows — supporting the ML pipeline end-to-end.

B.E. - Computer Science — Cloud Computing (IBM Specialization) - Chandigarh University (2017 - 2021)