Headline

Site Reliability & Platform Engineer @ Pluralsight | SRE · DevOps · Kubernetes (CKA) · AWS · Terraform · Kafka · Python · Linux | Distributed Systems, CI/CD & Incident Response

About

I'm a Site Reliability Engineer at Pluralsight working on Platform/DevOps engineering and production infrastructure. While plenty of engineers can ship a feature, my focus is the layer underneath — the systems that stay up, scale quietly, and never make it into a user's day. Reliability is an engineering requirement rather than an operations afterthought. Led the migration of legacy EC2 workloads to EKS — cutting compute cost 40% with near-zero deployment downtime — and built the observability layer (New Relic APM/NRQL, Grafana, SLI/SLO definitions, P95 latency alerting) that brought MTTR down 35%. My bar: if a failure mode isn't quarantined, observable, and rehearsed, the system isn't production-ready — it's just untested. Most of my work lives in three places: making infrastructure self-service (reusable Terraform modules, so teams provision without filing tickets), making delivery boring (GitLab CI/CD across 20+ microservices with Snyk scanning — 70% less deployment effort, 3× release frequency), and making failure legible (RCA, postmortems, and runbooks that actually get read). My Tech Stack: ▪ Cloud & Orchestration: AWS (EKS, EC2, RDS, S3, IAM, CloudWatch, Secrets Manager, SSM, Route 53), Kubernetes (CKA), Docker, Helm, IRSA, HPA, cluster autoscaling. ▪ IaC & Delivery: Terraform (reusable modules), GitLab CI/CD, GitHub Actions, everything-as-code provisioning. ▪ Observability & Reliability: Prometheus, Grafana, New Relic (APM, NRQL), CloudWatch; SLIs/SLOs/SLAs, error budgets, alerting, incident response, RCA. ▪ Languages & Data: Python, Bash/Shell, SQL, YAML; PostgreSQL (JSONB), Redis, PgBouncer, Kafka. ▪ Security & Governance: DevSecOps, Snyk, IAM least-privilege, secrets management, policy guardrails. ▪ AIOps: LangGraph, RAG over pgvector, Langfuse tracing, RAGAS evaluation in CI — building OmniSight, an LLM-powered incident triage platform. Performance engineering is where I enjoy myself most: a PgBouncer redesign that cut active DB connections 60% and improved p95 latency 25%, holding steady through 3× peak traffic. Same instinct that got me to Knight on LeetCode (500+ solved, top 4%) — find the constraint, remove it, measure it. 📩 Open to SRE / Platform / DevOps opportunities

Location

Bengaluru

AI stack

antigravity

claude code

claude

github copilot

mcp

claudecode

Work experience

Site Reliability Engineer, Cloud Platform at Pluralsight, Jul 2024 - Present · 2 yrs 3 mos, Bengaluru, Reliability & Observability: Built a full-stack observability framework with New Relic (APM, NRQL) and Grafana; defined SLIs/SLOs for critical services and configured automated alerting on P95 latency spikes, reducing Mean Time to Resolution (MTTR) by 35%. ▪ Incident & Escalation Management: Drove incident response and structured root-cause analysis with contributions to postmortems and runbooks, systematically reducing recurrence and improving overall service reliability. ▪ Kubernetes & Scalability: Led migration of legacy EC2 workloads to EKS using Docker, Helm, IRSA, HPA, and cluster autoscaling — achieving a 40% compute cost reduction (~$350k/year) with near-zero downtime during deployments. ▪ Infrastructure as Code & Self-Service: Authored and maintained reusable Terraform modules (ALB listener rules, Route 53, RDS, Redis, SQS/SNS, IAM roles/policies, Secrets Manager/SSM, CloudWatch alarms), enabling self-service, everything-as-code provisioning for development teams. ▪ CI/CD & DevSecOps Guardrails: Engineered GitLab CI/CD pipelines for 20+ microservices with automated unit/integration tests, Snyk container vulnerability scanning, and multi-stage deployments — cutting deployment effort by 70% and improving release frequency 3×.

Site Reliability Engineer Intern, Cloud Platform at Pluralsight, Jan 2024 - Jun 2024 · 6 mos, Bangalore Urban, ◦ Engineered cloud asset inventory system using AWS SSM, Lambda, and EventBridge to automate EC2 metadata collection across 200+ instances; built PostgreSQL database with JSONB columns and React dashboard that reduced audit prep time by 40% (30 → 18 hours quarterly). ◦ Built Jira automation platform using Node.js and REST API monitoring 15+ custom triggers with rule-based workflow execution; eliminated 50% of manual operations (15 hours/week saved), improved SLA compliance from 78% to 94%, and increased team velocity by 22%. ◦ Developed RESTful API backend using Express.js and PostgreSQL handling 500+ daily requests with P95 latency under 100ms; implemented normalized schema with proper indexing, JWT authentication, rate limiting, and achieved 99.8% availability.

Microsoft Engage Intern at Microsoft, Mar 2022 - Jun 2022 · 4 mos, Remote, ◦ Selected from 10,000+ applicants through competitive aptitude assessments and technical interviews to participate in Microsoft's flagship Engage mentorship program; built AI-powered Movie Recommendation System using collaborative filtering algorithms and machine learning that advanced to final evaluation rounds. ◦ Developed recommendation engine leveraging Python, scikit-learn, and content-based filtering techniques; implemented user preference analysis and similarity scoring algorithms achieving 85%+ recommendation accuracy on test datasets. ◦ Collaborated with Microsoft mentors on software design best practices, scalable system architecture, and Agile development methodologies; presented technical solution to engineering panel demonstrating problem-solving and product thinking capabilities.

Education

Bachelor of Technology - BTech, Computer Science, Rajiv Gandhi Institute of Petroleum Technology (RGIPT)

Public links

github: https://github.com/HarshUmpoxy

leetcode: https://leetcode.com/algoacer