About the Position

CloudThinker is an AgenticOps platform where AI agents operate cloud infrastructure alongside human teams. As a Forward Deployed Engineer you embed with our customers and make agentic operations real on their infrastructure: you encode their runbooks into skills, design the guardrails and approval policies the agents run under, graduate agents from notify to autonomous one trusted task at a time, and own the business outcome — not the ticket count. This is the role for a strong DevOps/SRE engineer ready to move up a layer: the agents take the execution, you take the judgment.

What You'll Do

• Embed directly with customer teams — onsite and remote — to deploy CloudThinker against their production AWS, Azure, and Kubernetes environments
• Translate the customer's tribal knowledge and runbooks into Workspace Skills the agents execute: cost optimization, incident response, patching, drift reconciliation, health checks, and reporting
• Design the trust ladder per environment and per task — what agents may do autonomously, what needs approval, and what evidence graduates a task to the next level
• Own guardrail and approval policy: scoped credentials, connection tiers, audit requirements, and compliance constraints
• Run the first incidents and cost runs side-by-side with the customer, then hand the encoded playbook to the agents and supervise
• Own the outcome numbers — cost reduced, MTTR cut, toil hours returned — and present them to customer engineering leadership
• Feed field reality back into Product and Engineering: missing skills, guardrail gaps, and the friction that separates a demo from a deployment

Requirements

• 3+ years hands-on DevOps, SRE, or cloud operations experience — you have carried a pager and made production changes under pressure
• Deep working knowledge of at least one major cloud (AWS preferred; Azure or GCP welcome) and Kubernetes — cost, IAM, networking, and observability included
• Runbook fluency: you have written the procedures, not just followed them, and you can decompose an operational task precisely enough for an agent to execute it
• Blast-radius judgment: you know which actions are safe to automate today, which need an approval gate, and how to tell the difference with evidence
• You drive AI agents and coding agents (CloudThinker, Claude Code, Codex, Cursor) as your primary leverage — specify, delegate, review, correct
• Customer-facing communication: you can walk a CTO through what changed and why in outcome language, in Vietnamese and English
• Consulting instinct without the consulting distance: you embed, you ship against real systems, and you stay until the number moves
• Nice to have: FinOps, security/compliance (SOC 2, PCI), or managed-services background; Terraform and IaC depth; incident command experience

What We Offer

• Be the forward edge of a new discipline — AgenticOps — with the platform team one desk away
• High ownership, real impact, and minimal bureaucracy
• Work with a strong technical team solving real customer problems
• Competitive compensation and flexible working model
• Long-term growth opportunities
• Hybrid working

What Success Looks Like