DevOps / SRE Engineer - AI Platform

The DevOps / SRE Engineer owns the operational substrate of an AI-native retail decisioning platform — infrastructure, CI / CD, observability, cost meter, and incident response for a system that runs production agents taking real business actions. The role builds on the enterprise Terraform standard, CI / CD spine, and FinOps tagging policy rather than reinventing parallel infrastructure. 

Remote candidates outside of Thailand are welcome to apply.

• Adopt the enterprise Terraform standard and module library for all platform infrastructure; author platform-specific modules where needed (agent runtime, vector DB, knowledge graph); run drift detection weekly. 

• Build platform-specific CI / CD pipelines on the enterprise spine — service deploys, agent deploys, eval-gate enforcement; integrate eval gates so no agent reaches production without eval pass. 

• Operate rollback orchestration with sub-15-minute recovery; quarterly game days. 

• Own the platform observability stack — OpenTelemetry, Langfuse for LLM traces, custom dashboards for per-agent cost. 

• Implement the per-agent cost meter end-to-end — token counts, vector queries, model inference, downstream LLM Gateway costs; surface cost data to the enterprise GenAI cost dashboard. 

• Stand up the platform on-call rotation; author runbooks for every production agent and service; lead incident response with measurable corrective actions. 

• Implement platform cost-tagging policy consistent with the enterprise standard (team, domain, environment, project, agent, suite, persona); report monthly to Cost Review. 

• Drive cost optimisation — right-sizing, caching, model routing decisions, reserved compute. 

• Bachelor's or Master's degree in Computer Science, Engineering, or a related discipline. 

• 5+ years SRE / DevOps with production ownership. 

• Terraform at scale — modules, state, drift, environment promotion. 

• CI / CD for data + ML / AI services (GitLab CI / CD or comparable). 

• Cloud platform (Azure preferred; AWS / GCP transferable). 

• Observability — OpenTelemetry, Langfuse (or comparable LLM traces), custom dashboards. 

• FinOps — tagging policies, attribution, optimisation. 

• Incident response — on-call, post-mortems, runbook authorship

• AI / agent platform SRE experience; cost-meter / chargeback systems built or operated. 

• Multi-cloud production experience; open-source contributions to IaC / observability tooling. 

• AI / ML / agent system observability instrumentation (LLM cost, agent cost, eval scores). 

• Vendor certifications such as HashiCorp Terraform Associate / Professional, Azure Solutions Architect Associate, or Databricks Data Engineer Professional.