Platform Reliability Engineer (Agentic AI)

Search Atlas

About The Position

The Mission: Building the Autonomous Nervous System Search Atlas is moving beyond suggestions to full execution. Our agent, Atlas Brain, handles SEO, AEO, Google Ads, and AI Content Generation autonomously—zero manual intervention. While Platform Engineers build self-service tools for developers, you ensure those tools enable autonomous AI execution with 99.99% reliability. You're not keeping dashboards alive; you're building the engine that allows an AI Agent to replace manual marketing execution. If the platform is reliable, the agent is unstoppable. What You Will Do: Architect the Autonomous Backbone Design and maintain the Kubernetes-based platform (EKS/GKE) that hosts Atlas Brain and its distributed agentic workers—handling millions of requests across SEO crawling, content generation, and ad optimization pipelines. Engineer for Zero-Touch Automate every aspect of infrastructure using Terraform, ArgoCD, and Go/Python. If you have to do it twice, it must be a script. Enable true "zero manual execution" at the infrastructure level. Scale Agentic Workflows Optimize ML inference pipelines for real-time agent decision-making Architect high-concurrency crawling systems that feed Atlas Brain's intelligence Ensure sub-second latency for agent task execution (SEO, Content, AI Builder) Handle high-frequency data pipelines: real-time bidding, SERP monitoring, content generation at scale Define Radical Reliability for AI Establish SLOs/SLIs specifically for AI execution success rates and agent task completion, not just "uptime." Design self-healing systems that preemptively resolve failures before they impact autonomous workflows. Observability for Agent Decisions Build distributed tracing and monitoring for complex agentic interactions—trace agent decision trees across SEO/AEO/Ads workflows, enabling rapid diagnosis of "why the agent made that choice." Implement OpenTelemetry, Prometheus, and Grafana for full visibility into autonomous execution. Safety & Guardrails Implement guardrails and safety controls for autonomous agent execution in marketing contexts—ensuring AI actions align with business rules, budget constraints, and compliance requirements. Design human-in-the-loop escalation paths for edge cases. Cost & Performance Governance Proactively optimize cloud spend and resource allocation (Karpenter/KEDA) as we scale to thousands of agencies. Balance performance with cost efficiency for unpredictable AI workloads. Why This Is Different Unlike traditional SRE roles focused on keeping services up, you're building the infrastructure that enables autonomous AI to execute business-critical marketing tasks. Every millisecond of latency you eliminate, every self-healing mechanism you deploy, directly impacts whether Atlas Brain can truly replace manual agency work. This is not traditional SRE—you're building the autonomous nervous system for AI execution. What Success Looks Like Atlas Brain executes millions of marketing tasks daily with <0.1% failure rate Zero infrastructure-related incidents requiring manual intervention during business hours Platform scales from hundreds to thousands of agency clients without reliability degradation Complete observability into agent behavior: "We know not just that the agent acted, but why" Ready to build the platform that makes autonomous marketing execution a reality?

Requirements

Experience: 6+ years in Platform Engineering, SRE, or Infrastructure roles within high-growth SaaS environments—with proven experience supporting AI/ML systems at scale.
Infrastructure as Code: Mastery of Terraform, ArgoCD, and GitOps workflows.
Container Orchestration: Expert-level Kubernetes (EKS/GKE) networking, scaling, security, and multi-tenancy patterns.
Hands-on experience with MLOps pipelines for autonomous agents
Model versioning and deployment strategies for continuous agent improvement
Prompt management and A/B testing of agent behaviors
Guardrails for safe tool execution and decision boundaries
Scaling AI inference services (LLMs, embeddings, classification models)
Languages: Proficiency in Python for building custom platform tools and automation.
Observability: Deep expertise in distributed tracing and monitoring for complex, event-driven systems—specifically for debugging AI agent decision chains.
Data-Intensive Systems: Experience with high-frequency data pipelines, web crawling at scale, real-time processing, and low-latency requirements.

Responsibilities

Design and maintain the Kubernetes-based platform (EKS/GKE) that hosts Atlas Brain and its distributed agentic workers—handling millions of requests across SEO crawling, content generation, and ad optimization pipelines.
Automate every aspect of infrastructure using Terraform, ArgoCD, and Go/Python.
Optimize ML inference pipelines for real-time agent decision-making
Architect high-concurrency crawling systems that feed Atlas Brain's intelligence
Ensure sub-second latency for agent task execution (SEO, Content, AI Builder)
Handle high-frequency data pipelines: real-time bidding, SERP monitoring, content generation at scale
Establish SLOs/SLIs specifically for AI execution success rates and agent task completion, not just "uptime."
Design self-healing systems that preemptively resolve failures before they impact autonomous workflows.
Build distributed tracing and monitoring for complex agentic interactions—trace agent decision trees across SEO/AEO/Ads workflows, enabling rapid diagnosis of "why the agent made that choice."
Implement OpenTelemetry, Prometheus, and Grafana for full visibility into autonomous execution.
Implement guardrails and safety controls for autonomous agent execution in marketing contexts—ensuring AI actions align with business rules, budget constraints, and compliance requirements.
Design human-in-the-loop escalation paths for edge cases.
Proactively optimize cloud spend and resource allocation (Karpenter/KEDA) as we scale to thousands of agencies.
Balance performance with cost efficiency for unpredictable AI workloads.