Build your Agentic Investment … | Partiful

Frontier AI labs, including Anthropic, OpenAI, and Kimi, report Finance Agent v2 as their flagship finance benchmark. It compares models on their ability to analyze periodic SEC filings from companies listed in the stock market.

IPO Finance Agent https://github.com/benstaf/ipoagent extended this approach to lengthier and more complex IPO S-1 filings, using the SpaceX IPO filing as a case study.

But SEC filing analysis is only one pod. An investment bank is a stack of specialized desks — diligence, valuation, capital markets, credit, comps, risk — each with its own workflow and judgment calls. The goal of the hackathon is to build that stack, agent by agent, benchmark by benchmark.

What you'll build - pick a pod, or connect a few: Diligence pod. Read-only work on filings that already exist — push IPO Finance Agent itself: harder domains, new filings, better retrieval. Valuation pod. Forward-looking, not retrospective — build a model from scratch (DCF, comps, LBO) and defend the assumptions, rather than answering questions about someone else's numbers. Comparable-company and secondary-market pricing (Forge Global, Nasdaq Private Market) as a callable agent. Capital markets / underwriting pod. Deal structuring under uncertainty — share class, pricing range, syndicate allocation — plus risk-factor interpretation for offering-readiness. No rubric answer, only tradeoffs. Credit pod. Pro forma leverage capacity and common-control recast implications for creditworthiness. Pitch /

Organizer: Mostapha Benhenda

Register