Fund-ly
Full-Stack DeveloperDec 2025 – PresentAI-assisted crowd-investment platform connecting emerging-market founders with investors — KYC, milestone-based funding and AI due-diligence scoring.
- Java 21
- Spring Boot
- Next.js
- TypeScript
- PostgreSQL
- Python
- LLM agents
- Chapa
Overview
Fund-ly connects entrepreneurs raising capital with investors looking to fund early-stage ventures. I built and deployed it end to end: a Spring Boot backend with JWT and Google OAuth2, RBAC and a Liquibase-managed PostgreSQL schema covering project listings, KYC/AML verification, NDAs, investment agreements, pledges and webhook-confirmed payments; a Python AI service for due diligence; and a Next.js front end with real-time messaging and admin tooling.
Results
Challenge
Entrepreneurs in emerging markets lacked a trusted platform to raise capital with proper investor verification, legal agreements and milestone-based fund disbursement.
Solution
A Spring Boot platform with verification pipelines (OCR, face match, liveness), NDAs and investment agreements, Chapa payments, milestone automation and WebSocket chat — plus a separate AI service for diligence and recommendations.
Key technical decision
Due diligence runs asynchronously in a separate Python AI service — an agent orchestrator with tools, RAG retrieval and structured JSON output, behind a rate limiter with retries and idempotency. The core platform stays fast and predictable, and model providers (OpenRouter, Ollama, Gemini) can be swapped without touching the Spring Boot API.
Highlights
- KYC/AML verification with OCR, face match, liveness checks and encrypted document storage
- NDAs, digital investment agreements, pledges and webhook-confirmed Chapa payments
- Milestone-based funding with task tracking, disputes and scheduled automation
- AI credibility scoring, summaries, risk flags and pgvector-backed investor recommendations
- Next.js front end with TanStack Query, Redux and real-time WebSocket / STOMP messaging