These notes are for hosts, moderators, introducers, and podcast producers booking Dr. Christan Grant. Each section stands on its own and drops into an opener as needed.
A computer scientist building AI for the writing systems, speech patterns, and shifting data distributions that sit outside the standard pipeline.
Dr. Christan Grant builds AI for data that current systems cannot yet process. He is the Arnold and Lisa Goldberg Rising Star Associate Professor at the University of Florida, where he directs the UF Data Studio under a single banner, Making the Invisible Computable. The record reflects it. A Best Paper Award at the 2025 IEEE ASRU workshop for work on audio large language models and child stuttering speech. The first computational pipeline for reading pre-colonial Mixtec codices, developed with six undergraduate coauthors at UF. Publications at KDD, ICDE, ACL, and EMNLP. Service on the VLDB 2026 Review Board, as an Area Chair for KDD 2026, and as a Distinguished PC member for ICDE 2026.
Christan Grant trained at the University of Florida’s database research lab, where he held an NSF Graduate Research Fellowship and an NSF LSAMP Bridge to Doctorate fellowship. After establishing his research program at the University of Oklahoma, he returned to Florida as an associate professor, and UF named him the Arnold and Lisa Goldberg Rising Star Associate Professor.
At UF he runs the Data Studio under a single idea, Making the Invisible Computable. His group names where AI systems fail when reality violates their training assumptions, then ships the system that handles the data anyway. Distributions drift and fairness guarantees erode. Children who stutter produce speech that audio LLMs cannot transcribe. Pre-colonial Mesoamerican scribes wrote in visual arrangements no NLP pipeline could read. Grant’s lab treats every one of these gaps as a research problem that deserves a working system at the end.
The community has taken notice. His paper on audio large language models and child stuttering speech won a Best Paper Award at the 2025 IEEE ASRU workshop. ICDE 2026 named him a Distinguished Program Committee Member. KDD 2026 named him an Area Chair. VLDB 2026 appointed him to the Review Board. Since 2022 he has completed 154 peer reviews across KDD, ICDE, ACL/ARR, VLDB, EMNLP, and others. His research draws support from NSF, Colgate-Palmolive, Amazon, FAA, USDA, and the Robert Wood Johnson Foundation, with over four million dollars in proposals under review at NSF, Schmidt Sciences, and Mozilla.
Few researchers work fluently across data management, NLP, and audio processing all at once. Grant does. He can speak to a KDD panel in the morning and to a linguist studying pre-Columbian writing systems in the afternoon, and the work holds up in both rooms.
Grant advises five PhD students at UF, mentors eighteen master’s students and a long list of undergraduate researchers, and co-chaired Yang Bai’s dissertation committee to his 2024 PhD. Yang is now at Meta. Outside the lab, Grant coaches an Integrated Product & Process Design team building accessibility software with Vispero Freedom Scientific, and he is coauthoring an applied data science textbook with Dr. Laura Cruz Castro.
Please welcome Dr. Christan Grant.
Pre-colonial Mesoamerican codices encode meaning through visual arrangement rather than phonetic symbols. Until recently, no NLP pipeline in the world could process a single page of them. Grant’s group built the first. Their 2024 AmericasNLP paper introduced finetuned vision models for Mixtec codex interpretation, coauthored with six undergraduate researchers. A 2025 paper at the Computational Humanities Research conference classified name-date figures in the codices, and a follow-up paper appears at LoResMT 2026. The MultiScript-30k dataset, under development in the Data Studio, aims to be the first cross-script translation benchmark of its kind. Grant has studied a historical Mixtec codex in person at the British Museum and shared results at the Computational Humanities Research conference in Luxembourg.
AI benchmarks assume clean text, standard speech, and stable distributions. Real data rarely cooperates. Grant’s lab works precisely where those assumptions break. The group builds fairness-aware learning for environments whose distributions change over time, NLP that reads writing systems arranged in visual rather than phonetic structure, and audio models that understand speech patterns outside the adult-standard range. The output is not only papers. It is systems that deploy, classifiers that run, datasets that ship, and interactive tools that practitioners install and use.
Dr. Grant is equally at ease on a data-mining panel, an NLP panel, and a responsible-AI panel. Open-ended questions work best. Invite him to explain the research identity behind the UF Data Studio, the Mixtec codex work, or where fairness research is heading once distributions stop sitting still.