I am a senior majoring in Computer Science at UC Irvine and a researcher. My work focuses on understanding why LLM-based systems fail and when their outputs can be trusted, applied to education, including automated grading, feedback systems, and learning analytics. I am passionate about building AI systems that make education more accessible, fair, and effective, and I aim to work at the intersection of artificial intelligence, learning science, and data-driven educational tools.
Vedant Khatri
B.S. Computer Science, University of California, Irvine, expected June 2027 · GPA 3.77
Researching AI in education: automated grading, feedback systems, and learning analytics.
khatriv1@uci.edu · GitHub · LinkedIn · Irvine, CA
About
Publications
Peer-reviewed conference proceedings
Xu, Z., Khatri, V., Dai, Y., Liu, X., Li, S., Zhang, X., & Yu, R. (2026). Enhancing LLM-Based Data Annotation with Error Decomposition. In Proceedings of LAK '26: 16th International Learning Analytics and Knowledge Conference (pp. 325–335). Best Paper Nominee. doi:10.1145/3785022.3785070
Peer-reviewed journal articles
Khatri, V., Sinha, A., & Kumari, R. (2022). A study of the impact of e-learning on the health of school going adolescents in Bareilly city: A cross-sectional study. Indian Journal of Forensic and Community Medicine, 9(3), 112–116.
Sinha, A., Khatri, V., & Nath, B. (2022). Relevance of health education to e-learning-associated problems among the school-going adolescents in Bareilly city: An interventional study. Journal of Family Medicine and Primary Care, 11, 6863–6868.
Under review
Khatri, V., Cusimano, A., Swiecki, Z., Xu, Z., Liu, X., & Yu, R. (2026). From Diagnosis to Redesign: Using Quantitative Ethnography to Improve Multi-Agent LLM Essay Scoring. Manuscript under review at the International Conference on Quantitative Ethnography (ICQE).
In preparation
Large Language Models for Educational Data Annotation at Scale. Manuscript in preparation.
Research
Data Science Research Assistant
Research on language models for educational data annotation, resulting in a paper accepted to LAK 2026 and nominated for the Best Paper Award. Designed and ran the LLM annotation experiments across four datasets, Bloom, MathDial, Uptake, and GUG, spanning five prompting strategies and two model versions, and contributed to the framework separating task-inherent ambiguity from model-based annotation error.
Also building an LLM classification pipeline that labels course assignments into 12 assignment-type categories, work now in preparation as a manuscript on LLM-based educational data annotation at scale. Serve Qwen3-235B-A22B (Mixture of Experts, FP8) through vLLM on GPU and run batch annotation with resumable processing across assignment sets. Iterated the prompt and codebook through six versions, adding negative gates, a title fast-path rule, and bright-line disambiguation between quiz and problem-set items, and evaluated each version using accuracy, confusion matrices, and error analysis against 300 human-labeled samples.
Research Collaborator
Developing the codebook and qualitative coding scheme for analyzing an LLM-based clinical-reasoning tutoring system, working with Yann Hicke. Coding real tutor-learner transcripts, and surfacing the finding that coding the tutor's moves alone cannot separate effective from ineffective tutoring, a result that is reshaping the study's direction. Also identified several corpus-level data-quality issues that would have corrupted automated coding.
Research Collaborator
Integrating two systems, PromptHive, which helps instructors generate educational questions through prompt engineering, and LLMRespondent, which estimates question difficulty using synthetic LLM respondents and Item Response Theory calibration, to build a difficulty filter for generated questions. Investigating why these difficulty estimates are unreliable, since a single LLM produces a narrow proficiency distribution and can confidently misrate questions, for example rating a simple plug-and-solve question as hard.
Undergraduate Researcher
Received UROP (Undergraduate Research Opportunities Program) recognition for a research project on the effects of e-learning on students, and presented findings at the UROP Symposium 2024.
Research Assistant
Cross-sectional and interventional studies examining adolescent health; data collection and statistical analysis.
Teaching
Course Manager
Run the operational side of one of UC Irvine's largest introductory computer science courses, with enrollment of roughly 175 to 340 students per term. Train and coordinate 3 to 4 undergraduate teaching assistants each term, manage roster add and drop syncing during the add and drop period, generate attendance sheets, run retake-ticket eligibility and delivery, generate secure per-student credentials for in-lab exams, and run grading for lab exams and zyBooks assignments. Operate, maintain, and extend the course's Python automation pipeline, which integrates the Canvas, zyBooks, and Google Sheets APIs, and add new features on request, such as lecture-attendance tracking and automated retake-ticket link delivery through Canvas.
Learning Assistant
Supported the C++ portion of the introductory programming sequence while serving concurrently as Course Manager for ICS 31. Ran office hours with one-on-one debugging support, assisted in lab sections, and helped prepare exam and assignment materials.
Grader
Graded programming assignments, projects, and exams, and held office hours.
Learning Assistant
Supported intermediate Python instruction. Ran weekly office hours, assisted in lab sections, and helped with assignment and exam materials.
Learning Assistant
Supported introductory Python instruction in the same course I later managed. Ran weekly office hours, assisted in lab sections, and helped with assignment and exam materials.
Service & Leadership
Co-author of paper accepted to the 16th International Learning Analytics and Knowledge Conference.
Conference volunteer, assisting with conference operations and attendee support.
Liaised with and advocated for international students, addressing concerns and organizing transition support events.
Head Boy (2022–2023) and House Captain (2020–2022), leading career fairs, cultural events, mentorship programs, and academic workshops for 800+ participants.
Skills
Python, C++, SQL · Pandas, NumPy, scikit-learn, TensorFlow, Keras · Flask · Git/GitHub, Jupyter, OpenCV, Linux, Bash
Awards & Funding
SURF (Summer Undergraduate Research Fellowship), UC Irvine. A competitively funded summer research fellowship.
UROP (Undergraduate Research Opportunities Program) recognition, UC Irvine.
Dean's Honor List, UC Irvine.
LAK 2026 Best Paper Nominee.