Technology News

Zhengxi Tan Served as Judge at AI Builders Foundation AI Agent Hackathon

Zhengxi Tan served as one of eight industry judges for the Snowflake x Beta Fund x EverMind Agent & Token Economy Hackathon on August 7, 2026, in Menlo Park, California. Organizers reported 750 applications and 121 admitted builders, creating a selective field for teams developing AI-agent products focused on intelligence cost, value, memory, and infrastructure. AI Builders Foundation's hackathon calendar listed the August 7 Snowflake Agent & Token Economy Hackathon in Menlo Park, while the Luma event page described a full-day program for builders working with Snowflake, Beta Fund, EverMind, and other event partners. Its public site identifies AI Builders Foundation as a 501(c)(3) nonprofit in Palo Alto focused on free hackathons, curriculum, community, and open-source tooling for AI builders. AI Builders Foundation describes its hackathons as free, public build events that move teams from project scope to live demos, a format that aligned with the August 7 event's build-and-present structure. The event placed agent economics at the center of the brief, with teams asked to consider token cost, persistent memory, context management, and whether AI systems can deliver value at sustainable cost. The event organized projects around three tracks: cost of intelligence, value of intelligence, and an open-category wildcard. The cost track focused on making AI cheaper to run through approaches such as prompt caching, model routing, batch inference, context compression, and spend dashboards. The value track asked teams to build products with demonstrated willingness to pay or a clear path toward revenue per user, while the wildcard track covered other work in the AI-token economy. The judging format combined live demos with submitted materials and rubric-based review. The Luma listing described a program that moved from opening sessions and build time into submissions, three-minute demos, audience voting, and awards. The event recap described asynchronous scoring through submitted materials and a scoring rubric. The published categories emphasized continuity, innovation, and pitch impact. The format gave each team a concise presentation window while leaving room for a separate review of submitted materials. AI Builders Foundation's judge announcement connected Tan's role to his experience in large-scale AI/ML model efficiency, inference optimization, and model-infrastructure co-design. Those areas match the technical questions raised by the hackathon, where teams were not only presenting agent concepts but also addressing the cost, memory, and infrastructure constraints that affect whether agent systems can run beyond a short demonstration. In that setting, infrastructure experience was relevant because cost and memory decisions shaped both the product demos and the stated competition tracks. Zhengxi Tan has more than seven years of experience across AI/ML model infrastructure, inference optimization, distributed systems, and financial technology. His work includes AI model inference, cost-to-serve and capacity optimization, and large-scale payment systems. He has also worked on advanced payment recovery systems at global scale. Tan holds a Bachelor of Science in Computer Science from the University of Michigan. Tan has also participated in startup evaluations from the founder's side. Earlier in 2026, he presented SpurPay, an AI-powered payment-recovery venture, at "One-Person Unicorn: The Soloist's Ascent | 2026 Pitch Day" in New York, where SpurPay received Second Place in the Audience Favorite Award category. The August 7 judging role placed Tan in a public event focused on whether AI-agent products can be economically and technically sustained. The assignment connected his AI infrastructure and financial-systems background with AI Builders Foundation's emphasis on practical products, live demos, and builder education.