Understand AI. Use it wisely. Train all the way to FDE.
Lumen is the continuous AI program from elementary literacy through Forward Deployed Engineer practice — specs before prompts, hallucination hunting, MCP integrations, agent orchestration, eval gates, and customer engagement loops. Standards-aligned (AI4K12 / CSTA), coach-led with Bea, labs that create aha moments in one period.
Demo password: lumen123
Standards-aligned
Mapped to AI4K12 Five Big Ideas and CSTA AI Learning Priorities (2025).
15 labs that teach judgment
From decision trees to MCP protocol studio, agent orchestration, and customer engagement loops — skills you use the same day.
Classroom → field
Teacher dashboards and PD for schools; FDE track for professional / applied AI training.
Trust by design
FERPA/COPPA-aware architecture. Student data is not used to train foundation models.
Pathways
One continuous sequence from elementary through FDE
Age-appropriate tracks that spiral the same big ideas deeper each year — then a professional FDE field track for applied AI engineering.
Kindergarten – Grade 2
AI Explorers
Young learners discover where AI appears in daily life, how humans create tools that sense and decide, and that technology can help or hurt depending on how people use it.
2 units · 6 full lessons authored
Grades 3–5
AI Builders
Students train simple models, inspect data, and practice fair design — building intuition for how learning systems work before formal algebra.
3 units · 7 full lessons authored
Grades 6–8
AI Investigators
Middle schoolers run mini ML pipelines, stress-test bias, analyze recommenders — and master the power unit AI as a Work Tool: specs before prompts, verification, privacy, and agency.
4 units · 11 full lessons authored
Grades 9–12
AI Foundations
High school pathway: representation, classical ML, generative systems, and civic impact — capped by Operator Skills (prompt systems, hallucination control, thresholds, personal AI playbook).
4 units · 14 full lessons authored
Undergraduate / dual enrollment
College AI Literacy & Technical Core
University-ready sequence: evaluation methodology, technical track, and Professional AI Practice (evals that matter, org prompt systems, go/no-go memos, operator portfolio).
3 units · 10 full lessons authored
Professional · Forward Deployed Engineer
FDE Field Engineer
Train the full Forward Deployed Engineer loop: turn ambiguous customer problems into production Claude systems — MCP integrations, cost-aware agents, eval gates, reliability, and portfolio-grade judgment.
5 units · 17 full lessons authored
Interactive labs
Students build mental models by doing
Deterministic, privacy-safe labs that run in the browser — no student PII sent to third-party model trainers.
Representation & Reasoning
Decision Tree Builder
Build yes/no rule maps that classify situations — the foundation of rule-based reasoning.
Machine Learning
Train a Classifier
Label examples, train a simple model, and read accuracy plus confusion counts.
Ethical Design
Bias Explorer
Swap training distributions and watch fairness and accuracy metrics move.
Machine Learning
Perception Lab
See how sensors capture signals and features differ from full understanding.
Humans & AI
Prompt & Evidence
Practice verification: claim → evidence → confidence for generative outputs.
Societal Impact
Recommender Toy
Adjust user histories and observe ranking feedback loops.
Ethical Design
Model Card Writer
Document purpose, data, metrics, limits, and ethical notes for a system.
Machine Learning
Neural Net Sandbox Lite
Toy network intuition: capacity, noise, and overfitting (HS/college).
Humans & AI
Hallucination Hunter
Classify AI outputs as supported, false, or unverifiable — and learn the tells of invented confidence.
Humans & AI
Prompt Architect
Build professional prompts: goal, constraints, format, critique loop, and refuse line.
Societal Impact
Privacy Trail
Rate real-world sharing scenarios and practice safer data hygiene with AI tools.
Ethical Design
System Tradeoff Studio
Set thresholds under uncertainty: false alarms vs missed harm, trust vs safety theater.
Humans & AI
MCP Protocol Studio
Design tools, resources, and auth boundaries for a Model Context Protocol server — the FDE integration surface.
Machine Learning
Agent Orchestration Lab
Wire cheap→expensive model cascades with deterministic escalation gates and cost budgets.
Societal Impact
Customer Engagement Loop
Run the FDE loop: ambiguous brief → scope → prototype plan → eval criteria → handoff memo.
For decision makers
What procurement committees actually ask for
Standards maps, privacy posture, classroom ops, and professional learning — not just a student login.
- Public standards explorer with lesson-level tags
- Role-based multi-tenant model (student, teacher, school, district)
- Teacher progress tables and unit assignments
- Educator PD modules to teach without being an ML engineer
- Privacy & security pages written for FERPA/COPPA conversations
- Pilot-friendly demo district with seeded classes