About
I’m Alex. I map systems, then make them boring in production.
I take production systems that nobody has mapped end to end and turn them into something a team can ship on. Five years of it: a stock exchange, a crypto exchange, a global reinsurer and a multi-tenant retail platform. Since May 2026 I have owned infrastructure workstreams on the AI platform of a global mobility company, where I migrated an ML estate of 51 pipelines and 1,500+ models off self-managed Kubeflow onto Databricks.
I have built platforms alone and inside DevOps teams: one client could not use a managed database, so I built the highly available PostgreSQL cluster behind their production system by myself, on a short deadline. I wrote the onboarding guide for new engineers on the ML platform. I have held production on-call for a trading platform, an internal platform and an AI platform, and I led the DevOps team on a multi-tenant retail platform.
- Based in
- Yerevan, Armenia
- Time zone
- UTC+42–3 h ahead of Central Europe, 8–9 h ahead of US Eastern
- Languages
- Armenian (native), English, Russian
How I take over a system
-
Map it
Inventory first: accounts, networks, identities, pipelines, data, and who owns each one. The 51-pipeline audit and the 3.19-million-event allowlist check both started here.
-
Make it safe
Close exposure and tighten access: private paths, least-privilege IAM, OIDC instead of long-lived keys, secrets in Vault. The last pass patched a critical kernel CVE across a GCP fleet through Terraform, moved CI off one engineer’s personal token and closed 80+ dependency alerts.
-
Make it observable
You cannot set an SLO without a measured baseline, so I measure first: the production ingest path sat at p50 8.1 s. Then the chain, from a SQL-warehouse detector through a Lambda webhook and an SNS topic into PagerDuty.
-
Turn it into code
Modules, reviewed plans and runbooks, so the team ships without asking me first. Teams run their own containers and never see the registry credentials, and a launch step cannot execute until 79 repositories have been checked for access.
Working rules
-
Everything as code
One OpenTofu and Terragrunt repository, applied through Atlantis after review. Private networking, Databricks governance and the serving path all live there.
-
Read-only before write
I start with an inventory and an audit: what exists, what is exposed, who has access, what is patched. That first pass found an allowlist of 37 entries with no owners, and 51 Kubeflow pipelines nobody had listed.
-
Every change through review
Plans are visible before they apply. I review the platform’s Terraform as a merge gate, and my own changes go through the same gate.
-
Decisions in writing
I write the team’s ADRs on security, recovery and platform limits. The recovery design with a one-hour RTO target is one of them, and it was approved.
AI agents are part of my daily toolchain, with guardrails: Claude Code skills and sub-agents for pull-request triage and Terraform-plan review, agents that check each claim against live repositories and tickets, and a person who makes every write. Their output goes through the same review as mine.
Experience
- May 2022–now
SRE / DevOps / MLOps Engineer
DataArt · Full-time
Platform, reliability and security work for clients in mobility, retail, trading, insurance and crypto, and for DataArt’s own engineering platform.
- AI platform, global mobility company · 2026–now Own infrastructure workstreams in a small platform team, all in OpenTofu, Terragrunt and Atlantis: the Kubeflow-to-Databricks migration, private DNS for PrivateLink-only Databricks, catalog governance, the serving SDK and its multi-region deploy path, SLO alerting and recovery design. I author the team ADRs, review Terraform changes as a merge gate and take part in production on-call.
- Multi-tenant e-commerce platform · 2026 Led the DevOps team. Authored onboarding changes for three retailers, all of which reached production; added an access-governance module and a deploy-wave step to a teammate-built provisioning platform; planned and ran a tenant migration that cut that tenant’s running cost by about 80%.
- Stock exchange platform · 2024–2026 EKS clusters as code with Terragrunt and Terraform, and workloads delivered onto them through Jenkins shared libraries, Helm and Liquibase. Level 3 production support and on-call for the trading platform.
- Reinsurance platform on Azure · 2022–2023 Designed the Azure platform in Terraform and moved hundreds of Java pipelines to reusable Azure DevOps templates, under regulated change management.
- Internal engineering platform, DataArt · 2022–2023 Migrated and maintained a platform where each team describes its services in one descriptor: built its Kubernetes cluster with Prometheus and Grafana monitoring, reachable only over private VDI, upgraded its Python engine and held on-call. With a front-end engineer, built a per-team dashboard of resources and logs, and an editor and publisher for descriptors.
- Highly available PostgreSQL · 2022–2023 Built the cluster alone, on a short deadline, for a client that could not use a managed database: Patroni, HAProxy and Keepalived, Ansible provisioning, backups, access controls and alerts on replication lag and failover.
- Crypto exchange · 2022–2023 Ansible and Jinja2 deployments, Keycloak identity, VPN DNS on PowerDNS and Prometheus monitoring.
- Oct–Nov 2026
Independent security auditor
Patient-facing healthcare application · Solo engagement
A read-only security review of a healthcare application before launch, then a re-audit two weeks later.
- Oct 2021–Feb 2022
DevOps Engineer
Freelance
Content-creator platform: WordPress on GKE, Datadog, GitLab CI templates, Jenkins RBAC, Amazon RDS and Cloud Run.
- Apr–Sep 2021
React Developer
Stone Valley LLC · Part-time
Front-end work in React and TypeScript before I moved into infrastructure.
Toolbox
- Cloud
- AWS (multi-account networking, PrivateLink, Route 53, Lambda, API Gateway, Step Functions, ECR), GCP, Azure
- Infrastructure as code
- Terraform, OpenTofu, Terragrunt, Atlantis, Ansible, CloudFormation
- Containers
- Kubernetes on EKS (clusters and workloads), GKE and AKS, Helm, Docker, Harbor
- CI/CD
- GitHub Actions, GitLab CI, Jenkins, Azure DevOps, release-please
- Reliability
- SLO and SLI design, Prometheus, Grafana, Alertmanager, CloudWatch, Datadog, PagerDuty, disaster-recovery design
- Security
- Cloud security audits, IAM and KMS, HashiCorp Vault, OIDC and workload identity federation, Keycloak, Trivy, distroless images
- ML platforms
- Databricks (Unity Catalog, Asset Bundles, cluster policies), MLflow, SageMaker, Kubeflow, Kafka, Spark Structured Streaming
- Data
- PostgreSQL (Patroni, HAProxy, Keepalived), MySQL, Delta Lake, BigQuery, Liquibase
- Languages
- Python, Bash, Go, TypeScript, Groovy
- AI-assisted work
- Claude Code with custom skills and sub-agents, MCP integrations with Jira, Slack and GitHub, automated pull-request and Terraform-plan review, local models through llama.cpp
Education
- 2017–2023
Bachelor’s degree, Information Technology
National Polytechnic University of Armenia
- 2021
DevOps bootcamp: CI/CD, containers, infrastructure as code, observability
Nana DevOps Bootcamp
Off duty
The opposite of on call.
- Building
- Conatus, a private journal for daily effort. Next.js inside Capacitor for iOS and Android. Every entry stays on the device, and there is no backend. On the App Store since November 2026.
- Watching
- Interstellar, True Detective, Eyes Wide Shut. What pulls me in: mystery, exploration and visual precision.
- Thinking about
- Consciousness, philosophy, science, and how things work.
- Drawn to
- Travel, art, brutalist editorial design and Japanese underground streetwear. Also retro games and early-internet details, which explains the small person on the front page.