Back to browse
AI trading platform with 34% returns (3 months) – seeking acquisition

AI trading platform with 34% returns (3 months) – seeking acquisition

by fkratzer·Mar 6, 2026·2 points·0 comments

AI Analysis

PassBold Bet

34% returns claimed; no methodology, backtesting warning, or independent audit shown.

Strengths
  • Real infrastructure (FastAPI, React, PostgreSQL, JAX PPO)
  • Portfolio health scoring and regime detection add practical value
  • Free beta with no credit card requirement lowers friction
Weaknesses
  • 3-month track record is statistically meaningless; cherry-picked bull market period unmentioned
  • No transparency on validation methodology, benchmark comparison, or survivorship bias controls
Category
Target Audience

Retail investors, robo-advisor platforms, fintech acquirers

Similar To

Wealthfront · Betterment · Robo-advisors

Post Description

I built an autonomous AI portfolio management system that runs on $300/month infrastructure.

Results: 34% returns over 3 months vs 7% S&P 500. Real capital, not backtesting.

Architecture:

Stage 1: LightGBM ranks 1,700+ stocks daily (80+ features) Stage 2: JAX PPO optimizes portfolio allocation Walk-forward validation, champion/challenger deployment Nightly retraining pipeline Recent additions:

Personal portfolio analysis with ML health scores Daily digest with Perplexity-summarized news Market regime detection (bull/bear + volatility) Stack: FastAPI, React, PostgreSQL, LightGBM, JAX/Flax, Claude API

Now seeking acquisition by brokerages, robo-advisors, or fintech platforms. Building this from scratch is 2-3 years of specialized quant work.

Live demo: https://acis-trading.com

Happy to discuss the ML architecture, validation methodology, or acquisition strategy.

Similar Projects