Apply on Employer Site

MagicSchool AI · 1 month ago

Senior Context Engineer, AI Systems

United States

Full-time

Remote

Mid, Senior Level

$160K/yr - $190K/yr

4+ years exp

MagicSchool AI is the premier generative AI platform for teachers, and they are seeking a Senior Context Engineer for AI Systems. In this role, you will architect and optimize AI agents' reasoning and memory management within educational workflows, ensuring coherent assistance for educators.

Artificial Intelligence (AI)E-LearningEdTechEducation

Responsibilities

Context Architecture & Token Optimization: Architect and implement adaptive context curation pipelines that determine what information enters each agent inference step, balancing comprehensiveness with the finite attention budget of LLMs to prevent context rot

Long-Horizon Task Management: Invent and operationalize memory compaction mechanisms and state management patterns that allow agents to maintain coherence across extended teaching workflows (lesson planning, differentiation, assessment creation)

Context Evaluation & Monitoring: Design evaluation pipelines that measure retrieval precision, token relevance, and reasoning coherence as context evolves across sessions. Work with the evaluations team on developing frameworks for measuring attention allocation and agent performance degradation

Just-in-Time Knowledge Retrieval: Build dynamic, runtime data fetching systems that enable agents to autonomously pull relevant curriculum content, student data, and educational resources exactly when needed, rather than pre-loading context with unnecessary information

Tool & Integration Design: Engineer token-efficient tool APIs and retrieval layers where each tool earns its place in the context window through clear utility and minimal overlap, with robust metadata to guide agent decision-making

Cross-Functional Collaboration: Partner with Product, Research, and Education teams to translate complex educational workflows into optimal context configurations, understanding which information signals matter most for different teaching scenarios

Model & Platform Integration: Collaborate with ML researchers and platform engineers to co-design architectures that integrate memory modules, retrieval adapters, and human-in-the-loop correction systems

Technical Mentorship: Guide engineers on context engineering patterns, teaching the shift from prompt-first thinking to holistic context management, token budget awareness, and dynamic information curation

Qualification

Distributed systemsRAG systemsKnowledge graphsSemantic searchPythonTypeScriptNode.jsPostgreSQLNextJSSupabaseLLM APIsModel Context ProtocolContext window optimizationAgent evaluationEducational technology familiarityTechnical mentorshipCross-functional collaboration

Required

4+ years building distributed systems

Hands-on experience with RAG systems, knowledge graphs, or semantic search platforms in production environments

Strong coding skills in Python, TypeScript/Node.js

Experience with our stack (TypeScript, Node.js, PostgreSQL, NextJS, Supabase) or similar

Proficiency with LLM APIs (OpenAI, Anthropic, etc.) and their context management patterns

Experience with Model Context Protocol (MCP), context window optimization for specific model families, or building context-aware agent frameworks

Understanding of or interest in how educational content is structured (standards, curricula, taxonomies), privacy requirements (FERPA/COPPA), and how context needs differ across teaching scenarios

Experience with agent evaluation, measuring context quality/relevance, or instrumentation for attention budget tracking

Knowledge of curriculum standards, learning progressions, or educational metadata schemas that inform context design