Staff Software Engineer - AI Infrastructure

MagicSchool AI

98d

About The Position

As a Staff Software Engineer for Agents Infrastructure, you'll architect and build the core systems that power MagicSchool's autonomous AI capabilities. You'll design the runtime, orchestration, and evaluation infrastructure that enables AI agents to plan, execute complex educational workflows, integrate with teaching tools, and operate safely in classroom environments. This is a high-impact IC role where you'll influence the technical foundation of how AI agents operate across MagicSchool's platform, mentor engineers on agent systems, and ensure our agentic capabilities scale reliably for millions of educators.

Requirements

• 7+ years building large-scale distributed systems with at least 2+ years in staff/senior roles.
• Hands-on experience building agent applications with tool calling, context engineering, multi-step reasoning, or autonomous workflow orchestration.
• Strong coding skills in Python, TypeScript/Node.js, or Go.
• Working knowledge of containers, virtualization, sandboxing, and security patterns for multi-tenant systems.
• Understanding of or interest in educational workflows, privacy requirements (FERPA/COPPA), and safety considerations for AI in classroom environments.
• Track record of architecting complex systems, mentoring engineers, and driving technical decisions that balance innovation with reliability.

Nice To Haves

• Experience with agent evaluation frameworks, preference optimization, or reinforcement learning for improving agent reliability.
• Familiarity with educational technology integrations, LMS systems, or EdTech compliance requirements.
• Background in ML/AI infrastructure, model serving, or AI safety systems.
• Experience with open protocols for tool connectivity (MCP-style connectors) or browser automation.
• Security hardening experience for untrusted code execution or multi-tenant educational environments.

Responsibilities

• Design and implement highly reliable agent orchestration systems, shared state management, memory systems, and multi-step workflow execution optimized for educational use cases.
• Build secure, sandboxed execution environments for agent actions with robust isolation, monitoring, and safety controls appropriate for educational contexts.
• Develop comprehensive evaluation frameworks for agent performance, safety assessments, A/B testing, and regression detection to ensure consistent quality in educational workflows.
• Create unified interfaces for multiple LLM providers, integrate with educational tool ecosystems, and build connector frameworks for curriculum data and learning management systems.
• Partner with Product, Research, and Education teams to translate complex educational workflows into reliable agent capabilities from prototype to production.
• Guide engineers on agent architecture patterns, evaluation methodologies, and best practices for building reliable autonomous systems.
• Optimize agent execution for cost, latency, and quality while maintaining the reliability standards required for educational environments.

Benefits

• Flexibility of working from home.
• Unlimited time off to empower our employees to manage their work-life balance.
• Choice of employer-paid health insurance plans, including dental and vision.
• Generous stock options, vested over 4 years.
• 401k match & monthly wellness stipend.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Number of Employees: 101-250 employees