Longaeva AI Hackathon

1. Overview

Why You Should Care

• Solve real problems that professional investors face every day — not a sanitized academic exercise
• Have your work evaluated, and potentially implemented, by a live investment team
• Compete for monetary prizes and the chance to spend time in the Longaeva office
• Network directly with Longaeva across the business from investment professionals to our Technology and Data Science teams

What Makes This Different

Most hackathons ask you to build something functional in 48 hours. We are asking for something harder and more meaningful: original thinking about how a business actually works. The challenge is not “build a model that predicts earnings.” It is “simulate the system that generates alpha as you define it and show your reasoning.”

2. Eligibility

Who May Participate

• Currently enrolled undergraduate or graduate students at an accredited US-based university
• Technical and interdisciplinary students encouraged: Backgrounds in AI/ML, computer science, engineering, math, data science, and investing are all encouraged
• All individuals must be enrolled at the time of registration and throughout the competition

3. The Challenge: "Businesses, Not Numbers"

Spirit of the Challenge

Don't only think of a company as a spreadsheet. Businesses change over time and what matters is not just what happened last quarter, but why it happened and what could happen next. Traditional investing models often rely too much on historical numbers and patterns. That can be useful, but it often breaks when the world changes. This challenge is about building something more predictive: a system that tries to represent how businesses actually work, so it can be run forward under different scenarios.

Objective

Build a model or system that can simulate likely outcomes for a company (or a small group of companies) that can be used to amplify portfolio returns. We encourage participants to go beyond traditional financial modeling methodologies and welcome contributions from technical and interdisciplinary students, including backgrounds in AI/ML, computer science, engineering, math, data science, and finance. Submissions should showcase creative methodologies, techniques, and UI/UX designs that could help investors make better trading decisions.

What We Mean by Simulation

We want you to build a model or system that can be run forward, not just describe the past. Components could include:

• Key business drivers such as demand, supply, capacity, backlogs, pricing, costs, churn risk, etc.
• Scenario analyses for critical fundamental drivers
• Probability distributions via stochastic runs, not point estimates

Deliverable

A working simulation for one or more companies.

Evidence & Inputs

• Anchor on real, observable signals
• Use a mosaic of alternative and unstructured data such as filings, transcripts, job postings, reviews, forums, etc. and not only clean or standard data feeds

Differentiation

Participants should avoid relying solely on standard approaches as the core solution (they may be used as inputs). Winning solutions will use an inventive, defensible approach that is hard to replicate.

Interpretability

Participants must clearly show which signals moved their outputs and why, with behavioral and causal logic to support their work.

Scale

Demonstrate how the framework could scale via repeatable pipelines, modular components, and consistent evaluation methodology.

4. Competition Structure — Three Stages