Math Education × AI
Jennifer Kleiman
I build AI tools for mathematics education, critically evaluate whether they are effective, and question what it means to be effective.Doctoral student at University of Georgia.

At a Glance
Affiliation
University of Georgia
Focus
AI in Mathematics Education
Methods
Design-based • Qualitative • Analytics
// Featured Projects
ArguAgent
AI-powered platform for scaffolding mathematical argumentation in small-group discussions. Analyzes student positioning responses to form optimal discussion groups based on argumentation sophistication.
Build-A-Bot Fraction Challenger
Students argue against AI chatbots that embody common fraction misconceptions. Three-bot progression system with achievement badges. Students learn by teaching bots why they're wrong.
Video Transcription Pipeline
Gemini-based pipeline for transcribing classroom videos with speaker diarization. Validated reliability framework with consensus algorithms for education research.
Agentic tool for verifying and formatting APA 7th edition references. Uses Google Gemini 2.0 with Search Grounding to detect hallucinations and correct citations.
Visual editor for creating Extended Toulmin argumentation diagrams used in mathematics education research. Supports argument components, teacher supports, and timeline annotations.
// Recent Activity
Upcoming • April 2026
AERA 2026, Los Angeles
"Can ChatGPT Help Analyze Collaborative Mathematical Argumentation?" with Yizhu Gao & Xiaoming Zhai
Systematic testing of whether prompt engineering enables LLMs to reliably label argument components in classroom dialogue, comparing AI performance against human coders on a 10-label coding scheme.
October 2025
NCTM 2025, Atlanta
"Empowering Students' Argumentation Skills Through AI: A Healthy Adversarial Approach"
Poster presentation on Build-A-Bot system where middle schoolers argue against AI chatbots embodying fraction misconceptions.
// Recent Writing
Coming December 2024
First post after comprehensive exam defense
Topics: What I learned, where the research is headed, and what's actually working in AI math ed