Artificial intelligence is often framed as a future that engineers will eventually step into. At Stevens, students are already there: presenting, testing, and debating how AI should work in the real world, not just in theory.
That momentum was on display at the iCNS AI Engineering and Science Symposium, where students, faculty, and industry researchers gathered to explore how artificial intelligence is moving from academic exploration to applied engineering practice. Hosted by the Department of Electrical and Computer Engineering, the event brought together campus researchers and external experts to examine both the opportunities and responsibilities that come with rapidly advancing AI technologies.
The symposium, formerly known as DuckAI, featured keynote talks alongside student poster sessions and live demonstrations highlighting projects developed during the Fall 2025 semester. Rather than focusing solely on conceptual models, many of the student presentations emphasized implementation — building tools, evaluating systems, and addressing emerging technical challenges in areas such as large language models (LLMs), safety, and optimization.
According to Min Song, Professor and Chair of Electrical and Computer Engineering, the evolution of project topics reflects how quickly student work in AI is maturing. Students are moving beyond foundational discussions of bias and model training toward improving and securing the newest AI architectures themselves.
The scale of participation underscored that shift. A total of 34 student teams presented posters and demonstrations, sharing research that connected classroom learning to real-world technical applications. The projects ranged from efficiency improvements in LLM deployment to questions of trust, transparency, and system-level reliability; issues increasingly central to how AI is adopted across industries.
Among the projects recognized during the symposium’s Elite Poster awards was LLM-Powered Agentic Systems and Applications, led by Ph.D. candidate Yupeng Cao, which explored AI agents capable of planning, reasoning and executing complex tasks beyond traditional chatbot behavior.
Second place went to DEMI: A Reinforcement Learning Agent that Embodies Collective Intelligence to Minimize Attrition, developed by Drishti Parekh, Eden Charles, and Meetkumar Gajera, applying reinforcement learning to address retention challenges.
A third recognized project, Trustworthy Models and Data by Sabbir Ujjal, focused on ensuring AI systems remain reliable, secure and free from bias.
Industry voices added another dimension to the conversation. Speakers from organizations such as opAIda, Prescient Design (Genentech), and NVIDIA discussed how AI is being applied in areas ranging from business strategy to drug discovery and autonomous systems, reinforcing the connection between academic research and professional practice.
For students, the symposium functioned not only as a research showcase but also as a professional environment, encouraging them to communicate technical ideas to industry leaders and engage in conversations about career pathways and emerging research directions. One external advisor described the event as offering a “comprehensive and optimistic view” of AI’s role in shaping the future.
What distinguished Stevens’ presence was its emphasis on responsibility alongside innovation. The event encouraged students to approach artificial intelligence with curiosity while recognizing the need for safety, trustworthiness, and human alignment as systems become more embedded in everyday life.
As AI continues to reshape engineering disciplines, the work presented at the symposium suggests that Stevens students are not simply learning how to use these technologies — they are helping define how they should be built, evaluated, and integrated. The future of AI may be global, but much of its groundwork is being laid in collaborative research spaces much closer to home.
