Senior AI Platform & Agentic Infrastructure Engineer
Not sure if you're a good fit?
Upload your resume and TixelJobs AI will compare it against Senior AI Platform & Agentic Infrastructure Engineer at OKX. Get a match score, missing keywords, and improvement tips before you apply.
Free preview · Your resume stays private
About the Role
Who We Are
About the Opportunity
OKX’s Internal Audit function has an early but working AI-native capability: a multi-agent platform (Hive Mind), agentic workflows, data pipelines, and AI-enabled tools that let a very small team punch far above its weight. Your job is not to rebuild it at today’s maturity. Your job is to take it to a regulator-grade production standard and well beyond, and to drive AI Enablement across the department. You own the foundation, the agent runtime, and the harness: a resilient cloud platform, the agentic runtime and evaluation harnesses that make agents trustworthy in a regulated setting, and the governed data and AI infrastructure everything else depends on. We hire on demonstrated building, not on claims or credentials. We expect you to be more capable than the hiring manager in your domain: you will co-own and challenge the tooling strategy, not just implement it. This is a two-person engineering team: you deploy, debug, and hotfix your counterpart’s stack when needed.
Environment
Cloud on AWS or GCP. Claude as the starting point in a deliberately multi-model architecture (Anthropic, OpenAI, Google), using the right model for each job, including the plugins and integrations you build. Google Workspace for reporting and evidence. JIRA for the audit team’s workflow. Lark for team communications, alert bots, and the corporate wiki. You inherit a working Google-native prototype estate (Apps Script web apps, Drive-synced automation, locally scheduled jobs, Claude Code agent tooling) and evolve it without breaking daily use. Infrastructure-as-code, CI/CD, and observability throughout. Treat this stack as the starting point, not a constraint: you build production-grade systems end to end on what exists today, and you are expected to propose, prove, and adopt better components as demand and capabilities evolve. Production-grade here means service levels sized for an internal assurance platform: board-cycle windows are sacred, recovery is measured in hours, and this is not a 24/7 pager culture.
In your first year
- Hive Mind runs in the cloud with HA, DR, SLOs, and audit logging that passes an internal controls review.
- A governed data foundation with provenance and lineage is live across multiple audit domains, integrated with OKX's group data infrastructure where it exists.
- An agentic runtime and harness with evaluation and red-teaming gates what reaches production.
- The hiring manager is out of the operational loop: no scheduled job runs on a personal machine, every system has a runbook and a non-founder owner. Decommissioning the founder’s laptop as infrastructure is a literal milestone. When these goals compete, the priority order is: keep the estate alive, then the cloud migration with observability, then the harness gating production, then the data foundation.
How we assess demonstrated building
We assess demonstrated building in ways that respect your time and your confidentiality obligations to current and former employers: a portfolio deep-dive (walk us through systems you built and kept running, at the level of detail your obligations allow; we want architecture, decisions, and trade-offs, never proprietary code, data, or documents), a short, time-capped build exercise on a synthetic problem unrelated to OKX's business (a small agentic workflow with an evaluation harness, used for assessment only and never put to use by OKX; the work remains yours) that you defend live, walking us through your design decisions and extending it on the spot, a systems-design session on taking a prototype estate to production, and references focused on whether you built and operated systems in production.
Trust and compliance
This role handles highly sensitive audit data at a global crypto exchange. Expect background checks, confidentiality obligations, and personal-trading and material-non-public-information
What You’ll Be Doing
- Inherit, operate, and progressively migrate the working prototype estate (Google Workspace–native automation across Apps Script, Drive, and the Docs/Sheets/Slides APIs; locally scheduled jobs; and Claude Code agent tooling) to the target platform without interrupting daily and board-cycle workflows. Working software wins arguments; migrate by strangling, not rewriting. Reuse and integrate with OKX's prevailing and evolving AI capabilities and data infrastructure, including enterprise-approved models and gateways, Model Context Protocol (MCP) servers and connectors, security tooling, and group data platforms, before building parallel capability.
- Re-architect Hive Mind into a resilient AWS or GCP platform with high availability (HA), disaster recovery (DR), defined service-level objectives (SLOs), and full observability, and own it in production.
- Build the agentic runtime and harness: orchestration, multi-model routing that sends each task to the right model, tool, MCP, and plugin integration across model providers, and the evaluation, red-team, and regression harnesses that grade agents before and after production, with evaluation gates, judge calibration, and cost controls that hold for each model, including prompt-injection and data-exfiltration threat modeling for agents that read untrusted content.
- Build the Responsible-AI and model-governance layer: hallucination, bias, and drift controls, output validation, guardrails, and complete logging. Extend the existing governance design (human-calibrated judge gates, golden-set regression, provenance registry) rather than replacing it. New model providers enter through
Ready to apply?
This job is active. Apply now to get in early.