An automated rule based reinforcement method for large language models that reward hacking and yields interpretable, dataset adaptive rules.
EigenAI
Help shape the future of AI and ML @ UTMIST
What is EigenAI?
September 20-21, 2025 | In-Person @ OISE, UofT
EigenAI is a UTMIST flagship conference introducing students to the world of AI, ML, software, and emerging technologies. Through panels and workshops covering both fundamental and advanced topics, participants gain hands-on experience and practical insights. This year’s theme is Mapping AI through the Multiverse, which invites students to journey through the many dimensions of AI, allowing them to explore the field from multiple perspectives and hear from professionals across diverse industries. Beyond technical talks and workshops, students have the opportunity to build their professional network and connect with industry leaders, academic professionals, and like-minded peers.
Schedule
Keynote Speaker
Sicong (Sheldon) Huang
PhD Student @ UofT | Co-Founder of UTMIST & for.ai
Speakers Session
Agentic AI and The Future of Everything
Karla Congson
CEO & CTO @ agentiiv
Responsible AI 101: Building Trust in Tomorrow's Technology
Rue Sriharsha
Senior Consultant @ KPMG
The Silent Layer: Privacy and Security at the Core of AI
Achin (Ace) Kulshrestha
Staff Software Engineer and Manager @ Google
Panel Speakers
Founders, Funders, and Future AI
Hassan Syed
Founder & CEO @ ASPIRE Pakistan
Maxim Miller
Partner & CPO @ QKFT
Jeremy Wang
Co-Founder & COO @ Ribbit
From Classroom to Lab: Undergraduate Research Journeys
Matthew Tamura
Research Assistant @ HCDS | APA
Jason Wang
AI Engineer Intern @ RBC Capital Markets
Sarah Walker
JD Candidate @ UofT Faculty of Law
Alden Arciaga
Laidlaw & Researcher @ UofT
Isaac Picov
AI Research Intern @ DevRev | prev Algoverse
This Year's Workshops
Mastering the Technical Interview: Strategies with Google
Building Applications with the Claude API
Accelerating Object Detection: Model Quantization & Deployment with NVIDIA Triton Server
AUTORULE: Reasoning Chain-of-thought Extracted Rule-based Rewards Improve Preference Learning
UTMIST Academic Department
Recommendation Systems: How Netflix Knows What You’ll Watch Next
Ax = λx
We bring together different eigenvectors - special kind of vectors often times with varying magnitudes and directions
With over 500+ participants, EigenAI is built for AI practitioners of all skill levels
Begineers: Learn fundamentals of AI/ML, meet peers and build your network within UTMIST
Enthusiasts: Explore professional networking and internships, learn about latest AI applications in industry
Experts: Connect with professors and industry leaders to understand ML applications in production and research
