Investment
Intelligence Engine

Technical Analyst — reading momentum
System nominal

Attention doesn't scale with headcount.

Senior advisors could give real, reasoned attention to maybe forty households each. Everyone else got a templated quarterly email and a rebalancing algorithm with no explanation attached.

Seven specialists, one accountable decision.

A technical, sentiment, and fundamental analyst each read the market independently. A bull and a bear researcher argue the position. A portfolio manager sizes it, a risk manager checks it — and a human signs off before a client ever sees it.

A short, auditable chain.

Agents

LangGraph

LLMs

Claude + FinBERT

Portfolio + Risk Engine

Black-Litterman

Human Review

Why, briefly.

  • LangGraph — every hand-off between specialists is an inspectable node, not a buried prompt.
  • Claude — reasoning long enough to explain itself in language a client can read.
  • Black-Litterman — blends every view so no single confident agent can dominate the portfolio.

Recorded results.

0

Portfolios advised

<4s

Avg. response time

0

Client retention

What's next.

Now

Live for paper trading, backtesting, and human-approved recommendations.

Next

Tax-loss harvesting that factors in each client's specific tax situation.

01 — Documentation

README

Complete project documentation.

02 — Source

Production Repository

Private source code.

Available during technical discussions.

Tech Stack

Frontend

Next.js

Backend

FastAPI

AI

LangGraph · Claude · FinBERT

Finance

Black-Litterman

Infrastructure

PostgreSQL · Redis

Interested in this project?

Let's discuss it — schedule a live walkthrough.