Head of AI Agent Systems

We Love AI Jobs; San Francisco (CA); On-site; USD 260,000 - 420,000; Full time

Job summary

Wonderschool seeks a Head of AI Agent Systems to transform its multi-agent system into a reliable platform that directly drives revenue. The role blends hands-on engineering with system architecture and leadership of a small team, ensuring adoption across the company.

You will own architecture, improve outputs, and tie agent performance to revenue growth while delivering end-to-end, production-grade workflows that reduce manual interventions.

Qualifications

• Hands-on engineering with system architecture leadership.
• Experience with AI agent systems, multi-agent coordination, and production reliability.
• Ability to design and implement evaluation and monitoring for agent performance.

Responsibilities

• Improve agent architecture, output quality, coordination, and reliability for production workflows.
• Build evaluation and reporting systems to track agent success, failures, and human intervention.
• Lead a small engineering team and drive agent adoption across company functions.

Skills

Multi-agent systems

OpenClaw

Claude Code

GBrain

LLMs

Agent evaluation

Workflow automation

DevOps

Tools

OpenClaw

Claude Code

GBrain

Job description

# Head of AI Agent SystemsWonderschoolSan FranciscoFull-timeAI tools:OpenClawClaude CodeGBrainUpload your resume once and Scout applies for youApply with ScoutShare this job:LinkedInXSave imageThe Head of AI Agent Systems will turn Wonderschool’s existing multi-agent system into a reliable platform tied to revenue and operational outcomes. The role combines hands-on engineering and system architecture with leadership of a small team and adoption efforts across the company.* Permanent* On-site### Skills and expertiseMulti-agent systemsOpenClawClaude CodeGBrainLLMsAgent evaluationWorkflow automationDevOps### Key responsibilities* Improve agent architecture, output quality, coordination, and reliability for production workflows.* Build evaluation and reporting systems to track agent success, failures, and human intervention.* Lead a small engineering team and drive agent adoption across company functions.## How this job comparesCompared with 19 open director machine learning jobs on October 10, 2026.Pay: This job does not publish its pay. The 8 open director machine learning jobs that publish US pay start at $260K a year on average.Similar open jobs: We Love AI Jobs lists 19 open director machine learning jobs at 18 companies.Remote work: This job is on site. 4 of the 19 director machine learning jobs (21%) are remote.AI in this role: High. AI is the core of the work: building, training or deploying AI systems. The role’s core responsibility is designing, building, evaluating, and operating Wonderschool’s multi-agent system. The posting names OpenClaw, Claude Code, and GBrain as components of that system.## Full descriptionAbout WonderschoolWonderschool builds software and systems that help businesses operate, grow, and generate revenue.We started in childcare, where we provide a platform that helps providers manage enrollment, operations, compliance, and payments. We also partner with governments to run large-scale childcare programs.We are a Series B company backed by Andreessen Horowitz, Goldman Sachs, Long Journey Ventures, and First Round Capital. The business is cash flow positive and expanding that position.We are building a unified platform that combines marketplace, SaaS, and government infrastructure. We are expanding into adjacent verticals over time, including healthcare, microschools, and other service-based businesses.A core part of our strategy is using AI to automate how work gets done across the company. This includes product development, operations, customer support, and go-to-market. We are also building systems to improve compliance, oversight, and fraud detection, particularly in government-funded programs.We have already deployed a multi-agent system using OpenClaw and our internal Claude Code system. These agents operate across product, engineering, design, data, and operations. Parts of the system work today. Much of it is inconsistent, hard to measure, and not reliable enough for production use.This role is responsible for fixing that and turning it into a system that drives revenue.Position SummaryThe Head of AI Agent Systems owns the design, performance, and evolution of Wonderschool’s multi-agent system.This role is responsible for turning a partially working agent system into a production-grade platform that directly drives revenue by improving how we acquire customers, onboard them, operate their businesses, and iterate on product.This is a hands-on role. You will work directly in the codebase, agent system, and infrastructure while leading a small team of engineers. You will be responsible for making the system work in practice and tying it to business outcomes.This role reports directly to the CEO.ResponsibilitiesDrive Revenue Through Agent Systems• Design and deploy agent-driven workflows that directly impact revenue generation• Improve conversion, onboarding speed, and provider revenue through automation• Partner with product and go-to-market to identify high-leverage revenue opportunities and implement them through agent systems• Use agents to accelerate experimentation on pricing, enrollment, and growth features• Tie agent performance to business outcomes such as revenue growth, conversion rates, and usageOwn Agent System Architecture and Performance• Own the architecture of the multi-agent system including OpenClaw, Claude Code, GBrain, and internal tooling• Improve agent output quality so outputs are usable without heavy human correction• Identify and fix system-level issues including poor outputs, coordination failures, memory loss, and context limitations• Design evaluation systems to measure success rates, failure modes, and reliabilityBuild a Reliable Multi-Agent Execution System• Redesign agent workflows so core processes execute reliably end-to-end• Ensure product development workflows from requirements through deploy can run through agents• Improve agent-to-agent coordination so tasks are completed without manual intervention• Introduce validation layers, retry logic, and clear ownership across workflowsDrive Adoption Across the Company• Enable engineering, product, design, go-to-market, and operations teams to use agents in their daily workflows