My dissertation investigates how purpose-built generative AI tools can support K-12 science teachers in linguistically diverse contexts. Focusing on four U.S. states with large multilingual learner populations (Arizona, California, New Mexico, and Texas), this research centers on educators working with multilingual learners.
It uses two custom chatbots — SciLing Coach (developed by me) and Possibility Thinking (PT) Bot (developed by Professor Ronald A. Beghetto) — to aid teachers in designing science lessons that are both rigorous (aligned with state science and language/literacy standards) and equitable. Teachers interacted with these AI tools in real time to co-create engaging and equitable science activities, leveraging creativity and inclusivity while meeting accountability and disciplinary literacy demands.
Unlike generic AI tools, these bots are customized for science instruction involving multilingual learners, with adapted personas, knowledge bases, and formats aligned to standards in AZ, CA, NM, and TX.
This work contributes to the field of disciplinary literacy for multilingual learners, recognizing their linguistic and cultural assets and helping teachers scaffold instruction without diluting academic content.
By showing how tailored AI tools can support teachers to meet both creative pedagogical goals and standards, this research offers empirical data that can guide state education departments in crafting policies about AI in classrooms.
Because the tools are designed with alignment to state science and language/literacy standards in mind, the findings can help policy makers see how AI can be integrated in ways that reinforce, rather than conflict with, regulatory expectations.
The study underscores the need for policies that include support for teacher training, clear guidelines for AI tool usage, and possibly policies around data privacy, equity, and ethical AI deployment.
Current Status: Data analysis ongoing
My latest AI project, developed under the Principled Innovation® framework, one of the nine design aspirations guiding ASU's evolution as a New American University, has led to the creation of a sophisticated AI application that is now live and available to the ASU community.
This tool is a comprehensive, multifaceted platform that integrates the moral, civic, intellectual, and performance-based character assets central to the PI framework. By embedding these principles directly into the innovation process, I designed it to help ASU faculty, students, staff, and external users generate ideas that are not only creative and impactful but also ethically grounded and socially responsible.
ASU Recognition: Recognized as the #1 most innovative university in the nation, ASU continues to lead in transforming higher education through technology and principled leadership. This AI application exemplifies that commitment, providing actionable guidance to ensure that every innovation reflects ASU's values of ethics, inclusivity, and forward-thinking solutions.
Current Status: Launched and actively serving the ASU community
As a Claude AI Builder Ambassador and founding leader of the Claude Builder Club (CBC) at Arizona State University, I'm spearheading efforts to build hands-on learning and community around AI tools and responsible innovation.
Organizing and co-hosting Anthropic-led workshops covering Claude Code, Building Agents, Writing Effective Tools for Agents, AI Research Salons, and Societal Impacts of AI.
Organizing student-led hackathons to solve real problems using AI, along with demo nights where students showcase their projects.
Facilitating regular discussion sessions on ethics, fairness, and responsible AI, as well as emerging topics, to ensure community awareness and thoughtful practice.
Collaborating with faculty, student organizations, and Anthropic to secure support, resources, and guest speakers.
Mentoring student members, building outreach to expand participation, and creating a space where students feel empowered to build, experiment, and lead in AI.
Current Status: Active Role
A comprehensive open-source resource designed to help K-12 educators harness the power of AI tools like ChatGPT, Claude, and Gemini to enhance their teaching practice. This practical guide bridges the gap between emerging AI technology and classroom application, making artificial intelligence accessible and useful for teachers at all tech comfort levels.
A signature approach for responsible AI integration focusing on Delegation, Description, Discernment, and Diligence
Ready-to-use prompts for Reading/Literacy, Mathematics, Science, Engineering, and STEM integration
Specialized strategies and prompts for supporting English language learners
Real classroom applications for lesson planning, assessment creation, and differentiation
Demystifies how AI works without technical jargon
Impact: This repository serves educators who want to save time on routine tasks while enhancing student engagement and learning outcomes. From instant worksheet differentiation to creating kinesthetic learning strategies, the guide provides immediate, practical value for busy teachers. The project emphasizes responsible AI use, ensuring educators maintain privacy standards and pedagogical integrity while leveraging these powerful tools.
Current Status: Actively maintained with community contributions welcome. Growing collection of teacher-tested prompts and strategies.
A comprehensive study examining a decade of elementary science teaching practices, analyzing how teachers integrated literacy into their science instruction across the United States. This research analyzed 386 implemented classroom lessons published in Science and Children journal from 2013–2022, providing insights into the real-world application of Next Generation Science Standards (NGSS) and Common Core State Standards (CCSS) in K–6 classrooms.
63% of lessons included some literacy practices, with speaking (40%) and vocabulary instruction (23%) being most common, while reading comprehension (13%) and listening skills appeared less frequently.
Inquiry-based activities were prominent (asking questions, analyzing data, constructing explanations), while mathematical reasoning (4%) and evidence-based argumentation (8%) were less frequently observed.
Support was evident in 24% of lessons, most often through non-linguistic strategies (62%) and instructional materials (53%). Teacher scaffolding (38%) and peer collaboration (9%) were used less frequently.
Texas and California showed higher levels of literacy integration despite their different standards adoption status, revealing notable variation across states.
This research illustrates the distance between policy aspirations and classroom realities, particularly in terms of educational equity. The findings suggest that practitioner resources may emphasize accessible entry points for literacy integration but provide fewer models for deeper disciplinary literacy development and inclusive instruction.
Combined inductive and deductive content analysis using established NGSS and CCSS frameworks to systematically code practitioner literature.
Current Status: Manuscript in preparation
New projects on the way...