Nandana Sheri takes first place at Stanford Hackathon
Nandana Sheri takes first place at Stanford Hackathon Heading link

Senior Nandana Sheri won Stanford University’s TreeHacks 2025, the university’s annual collegiate hackathon. She and her teammates Kenneth Yang, Ryan Vu, and Aryan Sharma, all from the University of Washington, took the top prize in the LiquidAI EdgeAI Track, one of four main tracks available to students.
The hackathon is highly competitive to attend–Sheri was one of 1,000 students invited to compete in the hackathon out of about 12,000 applicants. She was the only UIC student accepted. She met with her teammates before the three-day event on Instagram and Discord to develop the premise of their project.
“One of my teammates had an Apple Vision Pro, so we wanted to do something within the realm of virtual reality,” Sheri said. “We decided to focus on hardware capability, knowing that most of the other projects were likely to be software-focused.”
Their product, SurviveX, is an embedded AI assistant on a hands-free device with voice guidance to provide real-time survival assistance offline using Edge AI. A person can speak to their device and receive directions to say, start a fire, tend to a wound, or use the stars to navigate. Unlike ChatGPT, which provides all instructions at once, their design provides easier-to-follow step-by-step instructions. While designed for use with a Vision Pro, the assistant can also be used on an iPhone.
Sheri is very familiar with hackathons and worked on the first two iterations of UIC’s SparkHacks. She said this experience was different; in addition to seminars and workshops, engineers from the sponsor companies stayed during the 36-hour event, providing ongoing insights and feedback to the participants and helping them refine their pitches.
Sheri, who has a PC, worked on the machine learning model while her teammates worked on code on their Macs. Using the PC was fortuitous because they needed Windows and an Intel chip to generate their machine-learning file. Sadly, when she tried to increase the parameters used to train the machine learning model from one billion to three billion, her computer crashed. Her team approached Nvidia, a sponsor of another track, who provided her with $250 GPU credits to complete the work.
The organic communication throughout the event was a highlight for Sheri.
“I’ve never been at an event with that many companies without a resume in my hand, but they were still hiring–companies were like, ‘Oh, we’ll fast track your application,’” Sheri said. “The focus was on the technology.”
SurviveX was one of 57 projects in the EdgeAI track and won the top prize of $9,000 for Most Creative On-Device AI Deployment.
“Our project was a little more complicated, and not a lot of people were able to get their model working without WiFi,” Sheri said.
Sheri’s experience at UIC, and what’s next
Sheri has been active in her time at UIC: in addition to her involvement as a member of the organizing committee and the co-lead of SparkHacks, she is a curriculum design assistant for Introduction to Phonetics and Phonology (LING220; president of UIC’s Women in Computer Science, having served previously as the professional events chair and social outreach chair; and as a TA for Advanced Data Structures (CS 351), Data Structures (CS 251), Program Design I (CS111), Program Design II (CS 141), and Mathematical Foundations of Computing – Discrete Mathematics (CS 151), a course with over 500 students. She has held several internships and a research assistantship.
Sheri graduates in May and has applied to multiple graduate degree programs. She plans to focus on natural language processing and is very interested in conversational machine learning and linguistics. She hopes to eventually go into industry but wants to gain more knowledge through an advanced degree.
If she could offer any advice to her fellow students, it would be to participate in hackathons. She said UIC’s SparkHacks is beginner friendly and a great place to start.