First seen by Alion on Nov 21, 2026.
As a Senior AI Platform Engineer, you will design and ship cloud-native platform capabilities with agentic engineering as the default way of working. You apply agents, tools, and reusable workflows to accelerate Terraform, Kubernetes, and GitOps delivery-and improve that practice only where it creates measurable value. You do not build agent tooling for its own sake. You have already shipped agentic workflows in production; you will help build and scale that practice here-not learn it from scratch. You own multi-cloud foundations, security, and observability; raise the technical bar; coach peers; and drive cross-team change with high ownership and production judgment.
What You'll Do (Role Expectations):
- Use agentic engineering as the primary way to design, implement, review, test, and validate platform changes; measure cycle time, quality, and safety-not how many tools you introduce.
- Help shape shared agentic practices for the platform team-reusable workflows, skills, guardrails, and evals-so good patterns become defaults as the capability grows; do not invent tooling without a clear problem and measurable outcome.
- Extend or introduce agent/tool integrations only when they remove real toil or risk; retire or avoid integrations that do not improve outcomes.
- Apply proven orchestration patterns-tool calling, state, retries, human approval gates, evaluations, and guardrails (least privilege, injection resistance, audit, kill switches, rollback)-based on systems you have already run in production.
- Own reusable Terraform modules and lifecycle patterns across Azure (deep) plus GCP or AWS; production Kubernetes; GitOps delivery (Argo CD, Helm, GitHub Actions); and platform services (ingress/API gateway, mesh, Keycloak/OIDC, cert-manager, DNS, secrets, private networking).
- Engineer observability end to end (Prometheus, Grafana, Loki, Thanos, Alloy/OpenTelemetry or equivalent); deploy, scale, secure, and recover Kafka, Redis, and relational/document databases on the platform path.
- Trace and fix issues from application behavior through API, gateway/mesh, identity/certs, network, cluster, and cloud; partner with application teams so platform and agentic tooling reflect real service needs.
- Codify zero-trust and least-privilege controls and support FedRAMP/PCI DSS/SOC 2-style evidence with Security/GRC; drive adoption of agentic practices that work-pairing, shared standards, and measurable outcomes-not demos or experiments without a clear problem.
Who You Are (Success Profile):
- You have already shipped agentic engineering in production-not as a first-time learner. You use agents and tools to deliver platform outcomes, and you know when not to build another integration.
- You take ownership of outcomes end to end. You dig into issues with intent to resolve them-not hand off unclear problems or stop at the first blocker.
- You think in systems, not single tasks. You see how changes ripple across app, platform, security, and delivery paths before you ship.
- You work effectively across teams-application, platform, security, product-and drive change through clarity, pairing, and shared standards, not tickets alone.
- You raise the technical bar for agentic platform practice: standards, coaching, and adoption-not only your own productivity.
- You are strong at multi-cloud Kubernetes and Terraform platforms and use automation to make environments repeatable, secure, and maintainable.
What We're Looking For:
- 10+ years in software, platform, DevOps/SRE, or cloud engineering with substantial hands-on design and delivery.
- Prior, demonstrated production experience using agentic engineering workflows (tool-calling agents, reusable skills/workflows, evaluations/guardrails, and related integrations) to deliver platform or infrastructure outcomes. This role is not a place to try agentic development for the first time.
- Evidence of leading or strongly influencing how a team adopts agentic engineering-shared practices, coaching, or standards-not only personal use of AI assistants.
- Advanced Terraform and Kubernetes; production GitOps (Argo CD/Helm) and CI/CD; deep Azure plus GCP or AWS, with ability to transfer patterns across clouds.
- Strong Linux, networking, DNS, TLS/PKI, IAM/RBAC, and distributed-system troubleshooting; hands-on with several of API gateways, service mesh, Keycloak/OIDC, cert-manager, and secrets managers.
- Practical operations for Kafka, Redis, and relational/document databases in Kubernetes or managed services; observability depth with Prometheus/Grafana/Loki (or equivalent).
- Coding ability in Python or Go (or equivalent) with evidence of reusable automation or tooling others adopt.
- Proven ownership in a fast-paced environment: ship quickly while keeping security, review, and recoverability; track record of solving hard production issues and driving cross-team change with a systems view-not task-only execution.
What Will Make You Stand Out:
- Proven ability to leverage AI technologies and workflows to drive measurable operational efficiency (cycle time, toil, review lag, MTTR, safe automation rate).
- Experience defining shared agentic practices-skills, workflows, evals, guardrails-that other engineers adopted.
- Hands-on with modern agent frameworks or SDKs (e.g. Google ADK, LangGraph, LangChain, LlamaIndex, or equivalent) and durable workflow engines (e.g. Temporal) on real engineering delivery-not proofs of concept alone.
- Relevant AI / GenAI / cloud AI certification (e.g. Google, AWS, Azure, or equivalent vendor or industry AI credential).
- Certified Kubernetes Administrator (CKA) and/or HashiCorp Terraform Associate (or equivalent Terraform certification).
- Exposure to GitOps (Argo CD/Flux), distributed tracing (Jaeger/Tempo), FinOps, and security compliance frameworks (SOC 2, ISO 27001, FedRAMP-style evidence).
- Familiarity with LLM observability or multi-model gateway patterns when applied to engineering agents-not as a standalone ML platform career.
Education:
Bachelor degree or equivalent experience. Advanced degree preferred.
Skills
AWS, Azure, GCP, Kubernetes, Terraform

