Publications

Selected work in AI and computing education.

For the latest citation details and complete record, visit my Google Scholar profile.

2026

Confirming Correct, Missing the Rest: LLM Tutoring Agents Struggle Where Feedback Matters Most [Paper] [Code]

Tahreem Yasir, Wenbo Li, Sam Gilson, Sutapa Dey Tithi, Xiaoyi Tian, and Tiffany Barnes.

Paper in 21st Workshop on Innovative Use of NLP for Building Educational Applications (BEA), ACL 2026. Accepted.

When Verification Hurts: Asymmetric Effects of Multi-Agent Feedback in Logic Proof Tutoring [Paper]

Tahreem Yasir, Sutapa Dey Tithi, Benyamin Tabarsi, Dmitri Droujkov, Sam Gilson, Yasitha Rajapaksha, Xiaoyi Tian, DongKuan Xu, Tiffany Barnes, and Arun Ramesh.

arXiv preprint, 2026.

A Framework for LLM Integration in Secondary Education: Insights from Computing Teachers

Heidi Reichert, Tahreem Yasir, Malvika Satyavolu, and Tiffany Barnes.

Paper in International Conference on Artificial Intelligence in Education (AIED), 2026.

2025

Herald of Advancement, Disruption, or Both: A Systematic Literature Review on Student-Facing LLM Tools in Undergraduate Computing Education [Paper]

Benyamin Tabarsi, Tahreem Yasir, Heidi Reichert, Xiaoyi Tian, Shiva Gadireddy, Clara DiMarco, Daniel Briceno, and Tiffany Barnes.

TechRxiv preprint, 2025. Under review at Computers and Education: Artificial Intelligence.

First-Author Papers

Reverse chronological order

Tutoring evaluation pipeline

First AuthorAcceptedBEA @ ACL 2026

Confirming Correct, Missing the Rest: LLM Tutoring Agents Struggle Where Feedback Matters Most

Tahreem Yasir, Wenbo Li, Sam Gilson, Sutapa Dey Tithi, Xiaoyi Tian, Tiffany Barnes

21st Workshop on Innovative Use of NLP for Building Educational Applications, 2026

We benchmark seven LLM feedback agents across 10,836 solution–feedback pairs and identify systematic failures on valid alternatives and incorrect solutions.

arXiv Code

Multi-agent tutoring pipeline

First AuthorPreprint

When Verification Hurts: Asymmetric Effects of Multi-Agent Feedback in Logic Proof Tutoring

Tahreem Yasir, Sutapa Dey Tithi, Benyamin Tabarsi, Dmitri Droujkov, Sam Gilson, et al.

arXiv preprint, 2026

Verification improves outcomes when upstream feedback is error-prone but can degrade performance when feedback is already reliable.

arXiv

Co-Authored Papers

Collaborative work

PRISMA review flow

Co-AuthorPreprint

Herald of Advancement, Disruption, or Both: A Systematic Literature Review on Student-Facing LLM Tools in Undergraduate Computing Education

Benyamin Tabarsi, Tahreem Yasir, Heidi Reichert, Xiaoyi Tian, et al.

TechRxiv preprint, 2025

A PRISMA-guided review of 52 papers covering pedagogical roles, technical design, evaluation methods, and learner impacts.

Zenodo

Six-phase integration framework

Co-AuthorConference

A Framework for LLM Integration in Secondary Education: Insights from Computing Teachers

Heidi Reichert, Tahreem Yasir, Malvika Satyavolu, Tiffany Barnes

International Conference on Artificial Intelligence in Education, 2026

An empirically grounded framework of 27 implementation steps across six phases for integrating LLM tools in classrooms.