Inspiration
The inspiration occurs in the night, when pushing through becomes the key to the success of working together. Thinking of areas in which modern finance is still in progress and was dominated by McKinsey, we assessed the discipline of caring to find facts before the sunk costs destroy value. With a diverse group we assessed on how we can leverage our skills and found in private equity.
What it does
Athena is the Deal Operating System for private equityhes. We replace the entire fragmented stack — VDR, sourcing, diligence, IC memo — with one platform built around AI agents. We're the only EU-native, audit-by-design platform built for the regulatory environment European mid-market PE actually operates in.
How we built it
Building of building a full-stack, event-driven AI system
Development Layer
- Engineers build features in Next.js (v16) with React (v19) and TypeScript (v5)
- UI styled with Tailwind CSS (v4)
- Backend logic uses Prisma (v6) connected to Neon
- AI workflows orchestrated via Inngest + Gemini 3 through Vercel AI SDK
hexagonal principles with a DB and API infrastructure
- Vitest- Tests
- ESLint- Linting
- Prettier- Formatting
- Conventional Commits- Commit standard
Continuous Integration (CI)
- Install dependencies via pnpm (v10)
- Run full test suite (Vitest)
- Run lint + format checks
- Build Next.js app
-> Automatic deployment via Vercel -> Each release tagged (CI Release Latest tag) -> Preview environments generated per PR
Runtime Execution (Production)
Running on:
- Node.js (v24) with Live system flow:
- User uploads deal data
- Event triggered → Inngest workflow
- AI agents (Gemini) process tasks:
- parsing
- validation
- analysis
- Data stored in Neon (Postgres)
- Results returned via Next.js frontend
Challenges we ran into
time-structure
idea finding & execution
design challenge
find the right architectural design for the agencies
Log in or sign up for Devpost to join the conversation.