The New Website for YU Students
“Does anyone have the link for the shuttle app?”
“Where are my semester grades posted?”
“How does the new printing system work?”
YU students who haven’t muted their campus-wide WhatsApp groups are accustomed to seeing questions like these pop up on the daily, perhaps on the hour. FTOC behavior aside, this phenomenon is largely a result of the fact that YU’s student information is scattered across multiple websites without a centralized platform tying all the links together. Navigating the numerous links and trying to find the relevant information can often be a daunting task. Kfir Slonimski (SSSB ‘26) had an idea to optimize this system for himself and for the YU student population in general.
Slonimski grew up in a kibbutz in northern Israel and served in a combat unit in the IDF. He came to the United States to play soccer, and began studying in community college in Oklahoma and then in Kansas, before transferring to YU in fall 2023. He studies data analytics in YU, which influenced his interest in software projects and optimizing systems. Although he isn’t formally trained in programming or computer science, he — like many curious students of our generation — began exploring the limits of AI and what he could do with “vibe coding” (AI-assisted coding).
Slonimski built a website and named it YU AI Assistant, which unifies YU’s informational links while also containing a chat interface for users to ask relevant questions. It pulls live, real-time data from these websites, such as shuttle departure times, today’s dining menu, study room availability, Canvas deadlines and minyan times. He designed the app using a method called RAG (Retrieval Augmented Generation) which means that the AI references an authoritative source of data (in this case, the YU websites) before generating a response. The effect of this is that the app has a relatively low “temperature” — a measure of creativity — and is therefore more likely to give a response like “I don’t have information about this question” than it is to hallucinate and provide false information.
The development of this website took nearly three months, and Slonimski recalls spending many all-nighters in the library working on it. He shared that the discipline and work ethic he needed to complete this project were heavily shaped by his time serving in the army.
Slonimski presented the website as a final project for a course he took with Professor Yitzchak Rosenthal on how to build with and connect databases to AI, a course which he described as “one of the most beneficial [he] has taken so far.” Remarkably, nearly everything else he needed to develop this website (and the other GitHub projects he created over the years) was self-taught.
One notable feature of the website is an integration with YUTorah that — if developed further — has the potential to be an incredibly useful tool. Slonimski created an automation that transcribes shiurim from the YUTorah website. Once transcribed, these shiurim are added to a database so that if a user asks a question about them, the AI agent will locate the relevant chunks and respond with a clip from the shiur that most accurately answers the question. This clip is given with an exact timestamp so that the user can jump straight to the moment in the shiur that contains the answer.
So far, Slonimski has used this automation to transcribe over 150 shiurim of Rabbi Shay Schachter, one of the most prolific lecturers on YUTorah. A completion of this project would mean transcribing all of the shiurim on YUTorah in this manner, a goal which would be feasible on an increased budget that would allow him to compute more efficiently.
Slonimski initially gave access to a small group of students to use the app for a week, and he improved the system based on their feedback. Although the testing phase is complete, Slonimski continues to monitor usage patterns and adds features based on what students ask the most. He emphasized that students are welcome to contribute to the continually evolving website, whether it is class notes, course materials or other YU resources that could be added. Interested students studying computer science or data analytics are encouraged to collaborate on the project as well.
Speaking to Slonimski, it becomes clear that creativity and self-motivation — especially in this day and age — have the ability to propel someone even more than traditional academic success. There are powerful tools available at our fingertips, and a person with enough curiosity and interest can learn how to create something really amazing. In Slonimski’s words, “if there is something that you want to learn or to create, now is the time.”
Photo Caption: Slonimski presenting at the YU AI Club
Photo Credit: Kfir Slonimski
