We’re excited to invite you to this month’s Research Café featuring two graduate researchers whose work spans environmental stewardship and cutting-edge artificial intelligence. Joshua Kover will share his innovative approach to urban forestry management, redefining how communities can help shape sustainable, resilient urban ecosystems. Naina Chaturvedi will introduce Ignito, her multi-agent AI learning platform that personalizes education at scale and is already making waves in the AI-in-education community. Join us for an inspiring session highlighting the power of interdisciplinary research and the creativity driving graduate scholarship at the School of Graduate Studies.
Joshua Kover
I am a graduate student in the Rutgers Landscape Architecture department, working with the Rutgers Forestry Program and the Rutgers University Landscape Architect. Last year, I graduated from Rutgers School of Environmental and Biological Sciences as a George H Cook Scholar with a bachelor's in Landscape Architecture (BSLA) and a minor in Urban Forestry. My interests lie in ecological communities, urban forestry, and community engagement.
Fostering Democratic Urban Forestry Management
Urban forestry is understood to be the art and science of the management of trees in an urban context. In a traditional rural forestry context, 'stands' are delineated areas within the forest for management purposes based on soil, climate, and vegetation profiles. Each stand is managed as a discrete unit to meet specific objectives. However, defining stands within the urban environment is a current and ongoing conversation. Urban environments remain dynamic ecosystems, though they are highly engineered; the soil, microclimates, and vegetation profile are often not naturally occurring and are manipulated based on the surrounding built environment. Within this framework, the community represents the most dynamic component, inheriting both the ecosystem services and the potential burdens associated with urban trees. With community buy-in, there is a stronger sense of ownership and care for the trees. In addition, depending on the capacity and capability of that community, they can play a significant role in the survivability of urban trees. Thus, they must play a significant role in the urban forest conceptual design and management process. By synergizing a review of academic literature on proper management and design with an interview process of relevant social organizations that deal with community engagement, I can combine the art and the science of urban forestry to define urban stands. This research is intended to serve as an educational resource for industry professionals to help create more impactful, sustainable, and less costly urban forests.
Naina Chaturvedi
Naina is a PhD student researcher with publications at ACL, EMNLP, and workshops done at AAAI, ACM SIGCSE, where she advances work on LLMs and agentic AI for education. She is the founder of Ignito, a platform where Agentic Intelligence meets Education — adaptive learning platform powered by Multi-Agent AI Systems and Agentic Data Mining—an intelligent system for human learning that personalizes education at scale through real-world projects. Naina brings over six years of industry experience in data science and machine learning, ensuring that her research solutions are both robust and impact-driven. She is also a Coursera instructor and teaching assistant, dedicated to democratizing knowledge and empowering the next generation of technologists. She has worked as a system software engineer as well.
Adaptive Learning at Scale at Ignito: Multi-Model AI Agents for Personalized Education
We present Ignito, a novel web-based educational platform that leverages multi-agent agentic artificial intelligence and dynamic data mining to deliver personalized learning experiences across computer science, data science, and AI domains. Ignito's architecture employs specialized AI agents—including Code Analysis, System Design, and Research agents—that collaborate through a central orchestrator to provide comprehensive, domain-specific feedback and intelligent content curation. The platform integrates LLM-powered conversational assistants with vector database technology and semantic search capabilities to maintain educational focus while enabling natural interactions. Our agentic data mining system continuously discovers hidden learning patterns, predicts knowledge gaps, and triggers preemptive interventions through real-time behavioral analysis. The framework supports project-based learning with dynamic difficulty scaling, intelligent roadmap generation, and interactive system design exploration, all delivered through cloud-based infrastructure with fast response times. Building upon established research in AI-enhanced video learning and conversational educational assistants, Ignito represents a paradigm shift toward truly intelligent, adaptive web-based learning ecosystems that scale personalized instruction while maintaining pedagogical rigor. The platform's unified integration breaks down data silos, creating a single source of truth for comprehensive learning insights that inform both immediate interventions and long-term curriculum optimization.