LLM agents for quantitative workflows
Build and evaluate agents that can reason over financial documents, data pipelines, strategy research notes, and market-related tasks with careful benchmarks and failure analysis.
Research
Research Tracks
Our research program is designed to help members move from “I read the paper” to “I can reproduce, critique, extend, and explain the work.”
Evaluation, robustness, and interpretability
Study how LLMs fail, how to test model reliability, and how evaluation can be designed for high-stakes quantitative and research settings.
Market modeling and statistical learning
Explore statistical learning, time series, microstructure-inspired questions, and the boundary between classical quant methods and modern foundation models.
Retrieval and research infrastructure
Develop retrieval systems, experiment harnesses, clean datasets, and reproducible tools that make academic-style research easier to start and easier to verify.
Research Pipeline
Guided exposure before formal academia.
The club gives members a structured path toward academic research readiness and faculty outreach, without pretending that early students should already know how academia works.
Step 01
Read
Learn the background, identify the core contribution, and understand what problem the paper is actually solving.
Step 02
Reproduce
Implement a simplified version, check assumptions, and learn what details matter in real experiments.
Step 03
Extend
Design one small but meaningful extension: new benchmark, ablation, failure mode, dataset, or modeling idea.
Step 04
Communicate
Write clear research notes, present findings, and prepare for professor or lab conversations with evidence instead of vague enthusiasm.