Yerevan · On-siteSalary not disclosedfull-timeVerified recentlyStaff.am
This is a build role, not a ticket queue. Today, parts of the estate still run on legacy self-managed hosting and single-machine setups; access and security are managed informally; delivery practices vary by project.
Responsibilities
Build the governed GCP foundation — project and environment structure, IAM with least privilege, service accounts, secret management, logging, monitoring and alerting, budgets and cost discipline. GCP (BigQuery, Cloud Run) is the company cloud; you make it governable.
Migrate production workloads onto company cloud — including the AI agent runtime — off single-person and single-machine dependencies and legacy self-managed servers, without disrupting the live daily workflows the team depends on.
Own cloud security — one accurate, current access map (no unknown admins), secrets handling, audit logging, network posture, backup and disaster recovery with tested restores, and incident-response basics the team can actually follow.
Build the SDLC end to end — CI/CD on the company’s GitHub organizations, environment strategy (dev / staging / production), branch and release discipline, automated testing integration with QA, and AI-assisted code review in the pipeline. The team ships across four Shopify storefronts and internal services; you give…
Stand up and operate the AI platform’s infrastructure — the containerized agent runtime, a self-hosted open-source agent memory service (Docker + PostgreSQL on GCP, with a managed Cloud SQL path), and, as the platform matures, a model gateway for per-agent cost attribution and budgets.
Support the data platform’s operations — access, cost, and monitoring posture for the BigQuery warehouse, alongside the data engineer who owns its content.
Document as you go — infrastructure as code as the primary documentation, plus runbooks the rest of the team can operate from.
