A global hackathon for AI-native apps built with Jac.
Four days. Builders from every corner of the planet.
The standout projects — built over 4 days, online and worldwide, May 2026.
Anbuchelvan Ganesan
A multi-agent AI navigator that routes people in crisis — single mothers, undocumented families, elderly tenants — to the housing, food, healthcare, and legal aid they qualify for, in any language.
Soujanya Chatti
Your medications, safety-checked by AI agents that reason through your body's own chemistry — not a lookup table — flagging dangerous interactions and proposing safer alternatives.
Tony Y
A multi-agent council where seven AI specialists triage, debate, and cross-examine each case with evidence to surface the three most likely diagnoses.
Patrick Wang
The first autonomous prior-authorization agent that doesn't just submit claims — it fights every denial, generating ERISA-compliant appeal letters without human intervention.
Yash Patil
Five AI agents walk a Medicare claims knowledge graph to turn fraud signals into complete, cited investigations — not just risk scores — reaching 98% precision.
Ibrahim Pima
An agent that finds every recurring charge, judges which ones you actually use, and hands you the cancellation emails ready to send.
Viktor Nedev
An autonomous food-rescue platform that connects surplus food with shelters and volunteer drivers in real time, dispatching deliveries with AI agents.
Pranav Rajesh Krishnan
An autonomous OSINT agent that detects marine oil spills from live satellite imagery faster than any human, classifying slicks with trained ML models.
Ahmed Hammad
A full-stack Jac template enabling vehicles to communicate through an agentic protocol for fuel-efficient, safety-focused driving.
Kshitij Kumrawat
A graph-native life OS where six autonomous agents turn long-term goals into daily missions, coaching, and real-time progress tracking.
Harshith Nair, Anjika Singh
A matchmaking platform for builders — swipe through potential collaborators, match on skills and interests, and chat in real time to find your next teammate.
Shailesh Hawale
An LLM agent that investigates your dependency graph — finding CVEs, tracing transitive vulnerability chains, and generating validated fix plans in Jac.
Shashank Shekhar
A social layer on an agentic sleep platform — logs daily habits, computes personalized correlations, and nudges users toward better sleep.
Powering the future of AI programming
24 hours. 70 projects. One unforgettable weekend at U-M North Campus.
The projects that stood out — built in 24 hours at U-M North Campus, April 2026.
A K M Mahmudul Kabir, Abdullah Al Mahmud
A multi-agent AI platform where memory lives in a property graph, agents are graph walkers, and skills can be installed from the internet at runtime.
Aaron Sun, Ayush Madhav Kumar
Autonomous code defense with graph-native clarity. Monitors repositories for risky dependency changes and PR signals, with automated sandboxed testing and optional fix generation.
Adam Christley, Caeden Kidd, Amar Zecevic, Schultz
AI powered insider trading detection for prediction markets. Uses Tree of Thought reasoning to autonomously investigate manipulation, coordinated wallets, and suspicious trading patterns.
David Chong, Yanshengtian Dong, Shrimann Myneni
Detects fake wallet clusters in crypto airdrops using Jac's native graph Walker and AI reasoning. Builds a live wallet graph to distinguish coordinated fraud from legitimate activity.
Meron Demissie, Nitin Shankar Madhu, Chuka Ezeoke
An AI-powered mobile agent that hijacks your routine. Monitors device sensors and automatically performs contextual actions based on detected patterns and user-defined rules.
Krish Deshmukh, Kuldeep Debnath, Junna Park, Adam
An accessible learning toolkit for Jac combining a block-based visual interface, real-time graph visualization, and a VS Code Extension Suite with linting and debugging support.
Darsh Poddar
3D visualization of npm's single points of failure — with autonomous Jac agents that investigate any package's blast radius in real time across 217K+ vulnerability records.
Raj, Adwayan Goswami, Aryan Singh, Param Natu
A real-time AI assistant that listens, reads social cues, and suggests what to say instantly — for interviews, networking, and high-pressure conversations.
Anirudh Manjesh
Your codebase has an immune system. Corpus watches AI-generated code, catches breakage before it lands, and self-heals your project against your codebase graph.
Marcel Billingslea, Jana Hazimeh, Sudarshan Sridhar
Expose the infection. Build the immunity. A full-spectrum Bluetooth attack and defense system that weaponizes the trust graph — then inoculates it.
