Make learning visible
Define assessment signals and success metrics that distinguish performance, growth, participation, completion, and data quality.
Mathematics and learning-systems leader
Grant McCurdy
I build assessment systems, education data workflows, and AI-assisted instructional tools. My work connects mathematics leadership with practical evidence for teaching and learning.
Assessment and measurement, learning analytics, LMS/API automation, and human-reviewed AI.
Assessment & Measurement
Learning goals, success metrics, diagnostics, and reviewable evidence
Learning Analytics
SQL-backed reporting, statistical analysis, and decision support
AI-Enabled Systems
Human review, feedback, assessment, and operational workflows
Leadership & Implementation
Mathematics leadership, LMS/API workflows, and cross-team translation
Featured work / Logos Education Group
One assessment. A clearer view of the math program.
Explore a common mathematics assessment, its interactive department dashboard, and the report that connects findings to teaching priorities. All three use the same 30-question form; the report and dashboard share a fictional 500-student scenario.
Problems I Solve
From learning evidence to action
I design the connective systems that help people understand what students know, choose the next instructional move, and make organizational decisions from trustworthy evidence.
Connect evidence to instruction
Turn diagnostic and rubric evidence into reviewable feedback, remediation, and follow-up workflows without removing educator judgment.
Build decision-ready infrastructure
Translate LMS-shaped records into validated data models, analytics products, and explanations that different teams can use.
Selected projects
Assessment, analytics, and instructional tools
Working artifacts, the questions they answer, and the methods behind them.
Assessment measurementResearch prototype
Statistical Risk Modeling in R
- Problem
- Leaders need assessment-growth evidence that supports review priorities without turning model residuals into automated evaluations.
- System
- A public-safe R workflow searches model families, performs temporal validation and a locked holdout, and produces stakeholder-facing reports.
- Evidence
- The model card and verified reports expose decision guardrails, validity targets, and teacher, course, and section review signals.
- Limitations
- Some validity checks remain below target. The model supports review questions, not automated evaluations.
Assessment analyticsInteractive demonstration
Assessment Intelligence
- Problem
- A math department needs to turn assessment results into priorities for instruction and program review.
- System
- A common assessment, filterable dashboard, and department report share the same questions, synthetic data, and figures.
- Evidence
- Explore readiness across courses, Geometry class differences, track comparisons, and question-level teaching priorities.
- Limitations
- One designed fictional administration. It does not establish growth, causal impact, or placement suitability.
Assessment workflowAuthoring prototype
Assessment-to-Remediation Pipeline
- Problem
- Assessment authoring needs a controlled path from diagnostic design through review and offline LMS export.
- System
- The implemented slice validates 36 original math-readiness items, renders reviewer previews, and produces an offline Canvas New Quizzes payload.
- Evidence
- GPT-5.5 advisory reviews sit behind a human-review boundary before export; scoring, attempts, remediation, reassessment, and mastery reporting remain planned.
- Limitations
- The demonstrated workflow covers authoring, human review, and offline export. Scoring and remediation remain planned.
Working Evidence
Use the education systems
Featured interactive dashboard
Assessment Analytics
The Logos report as an interactive dashboard: courses, classes, questions, comparisons, and teaching priorities.
Supporting analytic workspace Education Data LabStructured chat surface for tables, summaries, charts, and cited analysis over a public-safe warehouse.
Supporting reviewer workflow Instructional AI Reviewer DemoInspect synthetic rubric evidence, draft feedback, review-required cases, approved language, and remediation planning.
Evidence & Method
Follow a finding back to its evidence
Portfolio summaries describe the problem and decision. Project pages show the system, validation, artifacts, and limitations. Repositories retain code, setup, tests, data contracts, model cards, and detailed architecture.
Review the evidence methodOpportunities
Better evidence for teaching and learning.
Through Logos Education Group, I bring assessment and reporting into a practical mathematics program review. This portfolio shows the analytical and technical work behind that practice.
I’m focused on learning engineering, AI learning systems, learning analytics, assessment and measurement systems, education data, and technical EdTech implementation. I also consider selective remote, part-time data and domain-expert AI evaluation work.