← Back to full-stack work

2025-12-12 / Built / AI & DeFi

YieldGuard

I built a chat interface for researching DeFi yield alongside volatility, gas costs, and risk preferences.

The problem

A yield number by itself leaves out a lot. I wanted to put pool data, market conditions, gas costs, and risk preferences in the same research workflow.

What I built

The Python and FastAPI backend runs statistical analysis and exposes an agent tool registry. The agent requests data as it reasons through a question. A Next.js interface provides chat, a dashboard, and wallet connection. Data comes from DeFiLlama, CoinGecko, and gas-data services.

What it shows

The project connects a language interface to structured analysis. It demonstrates that integration, not measured investment performance.

BUILT WITH

PythonFastAPINext.jsGroq