Published: 01/29/2026
Students Clarissa Aurelia Nahid Saputra, Pratham Patel, Pavan Kumar and Tejas Gupta secured first place in two tracks at NexHacks, one of the world’s largest hackathons, for their project WashedMCP, a semantic code search tool designed to optimize token efficiency in AI-powered coding assistants.
Modern AI coding assistants, such as Claude Code, consume significant numbers of tokens when searching through codebases, often processing hundreds of thousands of tokens to locate relevant code. This creates both cost and performance bottlenecks for developers working on large projects. WashedMCP uses semantic code search powered by a vector database to dramatically reduce token consumption, achieving a reduction from approximately 338,000 tokens to about 71,000 tokens — nearly an 80% improvement over conventional approaches.
“Attending hackathons like NexHacks has genuinely changed our perspective on what is possible in this field,” the students said. “Beyond the technical challenge, we had the opportunity to connect with students from universities across the country, learning how different programs approach computer science education and collaboration.”
Students also met venture capitalists and startup founders, gaining valuable insight into how industry leaders are thinking about the future of AI and its impact on computer science. Conversations with the organizing team taught the students a great deal about event management and coordination at scale; skills that have already proven useful in developing the Programming Club at Gannon University.
With mentorship from computer and information science (CIS) professors, including John Coffman, the students said they are grateful for the technical foundation and problem-solving skills developed through their coursework and involvement in the CIS department, which directly contributed to their success.