Open-source Lumina Capital: a 7-persona LangGraph AI hedge fund engine with Qlib Alpha158
OpenComplaint524 · reddit · 2026-10-22
A Redditor open-sourced Lumina Capital v6.0.0, an autonomous multi-agent quant hedge fund engine that strictly separates LLM probabilistic reasoning from deterministic quant computation to avoid hallucinations from feeding raw charts to LLMs.
Key architecture:
- 7-Persona Anti-Sycophancy Council (LangGraph): Buffett, Soros, Burry, Cathie Wood, Graham, O'Neil and CIS personas analyze in parallel, then hold adversarial bull-vs-bear debates; a dedicated CRO agent holds absolute veto over tail risks
- Pure Python Qlib Alpha158 + LightGBM: fully vectorized implementation of Microsoft Qlib's 158 quant factors with a LightGBM meta-decision filter
- Google TimesFM: 5-day probabilistic fan charts with 80% confidence bands and crash probability
- 10-second sentinel thread: monitors dynamic ATR stop-loss floors and executes physical exits without waiting for LLM loops
- Neo4j 3D causal knowledge graph + GraphRAG: supply chain choke-point surveillance and episodic memory
Ships with a Streamlit dashboard, released under CC BY-NC 4.0 for research/education/simulation. GitHub: ramuza0113-sketch/lumina-ai-hedge-fund.
Related event: Open-Source Multi-Agent Hedge Fund Engine Lumina Capital v6 Released(3 posts)→
More from coding & agent
- Red Team Pass prompt: make the model critique and fix its own output — thisguyknowsai · 2026-10-22
- Assumption Audit prompt: list assumptions first, then confirm before answering — thisguyknowsai · 2026-10-22
- Confidence Scorer prompt: make models label every claim HIGH/MEDIUM/LOW — thisguyknowsai · 2026-10-22
- Leaked NVIDIA internal prompting playbook claims 41% hallucination cut — thisguyknowsai · 2026-10-22
- Dev argues voice or a mechanical keyboard are the only good ways to talk to agents — SimonBalmain · 2026-10-22
- Where's a pixel-perfect Claude Code-grade terminal UI for open models? — victormustar · 2026-10-22