Building an AI that analyzes stocks like Warren Buffett

David Kim — Sat, 04 Apr 2026 04:59:10 +0000

Most financial tools give you data.

But investors don't make decisions using raw numbers alone — they interpret them through a framework.

Warren Buffett thinks about moat and intrinsic value.

Ray Dalio thinks about macro cycles.

Charlie Munger thinks about avoiding mistakes.

So I asked a simple question:

What if an AI could apply these investment philosophies automatically?

That idea became Wallstreet-AI, an open-source agentic financial analysis system that combines structured data pipelines with LLM reasoning.

GitHub:
https://github.com/davidkim205/wallstreet-ai

What is Wallstreet-AI?

Wallstreet-AI is an LLM-powered financial analysis assistant that converts natural language questions into structured investment reports.

Instead of manually gathering:

financial statements
technical indicators
earnings summaries
market news

you simply ask:

"What would Warren Buffett think about Apple in a high interest rate environment?"

The system builds a full reasoning pipeline and generates a structured report.

Key Idea: Persona-driven financial reasoning

Different investors interpret the same data differently.

Example:

Buffett focuses on:

business quality
durable moat
long-term cash flow

Dalio focuses on:

macro regime changes
interest rate cycles
portfolio diversification

Wallstreet-AI lets the same dataset produce multiple interpretations depending on the persona applied.

System Architecture

The project is designed as an agent workflow:

Natural language intent parsing
Tool routing based on analysis type
Data collection via APIs
News enrichment using RSS scraping
LLM synthesis
Streaming output via SSE
Structured logging for reproducibility

Pipeline:

User Query

→ Intent Parser

→ Tool Router

→ Data Collection

→ LLM Generation

→ Structured Report

Example Question

"What would Warren Buffett think about Microsoft today?"

The pipeline automatically:

detects ticker
determines analysis type
gathers market data
generates structured reasoning
streams the output in real time

Live Demo

You can try it instantly:

HuggingFace Spaces:
https://huggingface.co/spaces/davidkim205/wallstreet-ai

Google Colab:
https://colab.research.google.com/drive/13rWqKpAgJMytsztSt_spsws4HPWQZPnc

Future ideas

Possible extensions:

portfolio optimization personas
backtesting integration
vector database memory
evaluation benchmarks for financial reasoning
multi-agent debate between investor personas

Feedback welcome

I would appreciate feedback on:

agent architecture design
persona prompting approaches
evaluation methodology for financial reasoning
additional financial datasets