Run of show
Overview
2:50 – 3:15 PM · Research Slam
Research Slam: Lab to Market
Preference Modeling for LLM Alignment under Heterogeneity
Making LLM post-training well aligned with human goals.
What if AI Fails Exactly When it Matters?
We ask how to better emulate rare, extreme events when high-resolution training data are scarce by using the geometry of nudged ensembles to identify local instabilities and inform existing surrogate models non-intrusively.
Reproducing ELEPHANT: Judge-Dependence and Model Deprecation in an LLM Sycophancy Benchmark
Reproducing the ELEPHANT LLM sycophancy benchmark, we find its results partly hold but depend on judge and evaluation settings, and that such benchmarks face a short "reproduction window" before the models they evaluate are retired.
Statistical Frameworks for Reliable Generation
Controlling risk while preserving utility in language model generation.
Through the Eyes of the Citizen
Individualized AI for trustworthy urban evaluation
Judges
3:15 – 3:30 PM
Hackathon: Dorm to Market
Presenters
Experative
Experative builds the memory and decision-making layer for autonomous scientific discovery, enabling self-driving laboratories to learn from every experiment and continuously improve what they do next.
Daniel HofmannCo-Founder & CEO
Tom BurkartCo-Founder, Head of Product
FirstSignal
FirstSignal is an AI-powered venture sourcing and diligence platform that discovers promising founders early and uses evidence-grounded multi-agent analysis to evaluate them against an investment thesis.
Mingjia ZhangHackathon Finalist
Haggl
Patients who challenge a medical bill usually win. Most who don't assume it's pointless. Haggl finds the errors and savings, then handles the calls, emails, and letters to win them.
Jay VachhaniCo-Founder & CEO
Hamza MalikCo-Founder & CTO
Susy LiuCo-Founder & CPO
Other finalists
AMIRA
AMIRA is an AI-powered evidence navigation platform for women’s hormonal health that helps clinicians and patients understand what the evidence says, where it conflicts, and what remains unknown.
MD S. RanaFounder
Anaphora Labs
Anaphora Labs is the AI-native operating architect that maps, redesigns, and automates how enterprises run.
Samuel ChenCEO
Emiliano Camacho-AlmanzaCTO
Yaphet LemiesaChief Research Officer
Antemyth
AI that runs your company’s IT end to end.
GPU
Developed GPU architecture that accelerates RISC-V and C programs with warp cores and vector lanes, achieving high performance on parallel workloads such as graphics, physics simulations, and AI model inference—all in 24 hours.
Yolanda HuHackathon Finalist
Laufwerk
Your enterprise’s expertise should scale like software. Laufwerk provides a self-hosted platform to build and run AI systems that make it possible.
Ralf BoltshauserCo-Founder
Vera Labs
At the intersection of AI & Real World Execution
Ana RathFounder & CEO
Tanmay Gupta
3:30 – 3:45 PM
Startup Showcase
Sona8
Building the context layer for enterprise AI through voice agents that survey employees.
Anton HantelCo-Founder & CEO
Waku.one
Waku.one captures decisions as your team makes them, so agents never ask twice and never act on stale ones.
Shen Sean ChenFounder & CEO
ThoughtForge AI
ThoughtForge is building an AI control platform that enables robots to adapt in real time to enable higher uptime and faster deployment across tasks and environments.
Matthew BrownFounder & CEO
Vound
A robotics startup building data and deployment infrastructure for physical AI
Aly SiessaCo-Founder
- Youssif SiessaCTO
- Adham
Hosted by
Eesha Khare
Mariana Minerals
Staff Product Manager
Linn Bieske
MITAI 2026
Hackathon Lead
Research Slam · Hackathon · Startup Pitches
Judging
Celebrating the ideas, research, and demos shaping what comes next.