Research
Our research spans the theoretical foundations of machine learning, applied ML systems, and a dedicated quantum ML division — held together by a shared commitment to rigorous, collaborative work.
Project Alpha investigates a specific hypothesis in LLM optimization: that a large reasoning model's chain-of-thought, captured at an early stopping point, can be handed off to a significantly smaller model — and that the smaller model, guided by the partial thinking trace, can arrive at the correct answer at a fraction of the inference cost.
We benchmark frontier and open-weight models across a variety of challenging evaluations to identify where this handoff works, where it breaks, and what properties of the thinking trace matter most for successful transfer. The goal is a practical framework for high-accuracy, low-cost inference that doesn't require running a massive model to completion on every query.
Research Streams
Each stream meets regularly, runs collaborative projects, and produces technical output for the community.
Our QML division explores how quantum computing intersects with machine learning. We investigate foundational questions about quantum approaches to learning tasks and study the current state of quantum hardware, maintaining an honest perspective on where quantum methods may offer real advantages.
We study the mathematical and theoretical foundations of modern machine learning — how models learn, why they generalize, and what their fundamental limits are. Members read frontier papers and engage with topics ranging from optimization theory to large language models and reasoning systems.
Members build real ML systems and develop hands-on skills. Projects span training custom models, building end-to-end applications, and designing evaluation pipelines — gaining practical experience with modern frameworks and deployment.
How It Works
MIC research follows a deliberate progression designed to take members from paper consumer to paper producer.
New members start with the Theory Reading Group, working through textbook-level material and seminal papers to build shared vocabulary.
Once fluent in fundamentals, members join a focused research stream. They read frontier papers, attend weekly meetings, and contribute to discussions.
Members are matched with active projects. They contribute code, analysis, literature review, or experiments — with mentorship from research leads.
Mature contributors produce technical reports, workshop submissions, or open-source tools. Research leads guide writing, reviewing, and submission.
Collaborate
We actively seek partnerships with faculty, external research groups, and industry labs.
Propose a supervised project for MIC members, co-run a reading group, or offer mentorship in exchange for motivated research assistants.
Sponsor a project, provide cloud compute credits, or co-design a challenge problem. We produce rigorous, citable technical output.
Give a guest lecture, co-author a technical report, or open a collaborative research thread. We welcome connections across institutions.
Join a stream, pick up an open GitHub issue, or pitch a new project. All levels of experience are welcome.