[Discussion] Applying financial LLMs to China A-share factor research

[mykernel]

Hi FinGPT team,

Thanks for building such an important open-source financial LLM project. FinGPT’s focus on financial sentiment, financial data pipelines, instruction tuning, and robo-advisor style applications is very inspiring.

I am building a related application-layer project: Factor Lab.

Factor Lab focuses on China A-shares. The core workflow is:

• build factor-based and event-driven stock pools;
• identify profit-gap / earnings-related candidates;
• use multiple AI roles to generate roundtable-style equity research reports;
• keep historical stock pool and report results for review.

Compared with FinGPT, Factor Lab is not trying to train a financial LLM. It is more of an applied research workstation where financial LLMs can be used inside a concrete A-share factor workflow.

I would love to discuss:

• how financial LLMs can be evaluated in real investment research tasks;
• whether factor scores and historical stock-pool results can become useful feedback data;
• how Chinese market data, announcements, and earnings events can be better represented for LLM analysis;
• whether FinGPT-style models can support explainable A-share research reports.

Project link: https://www.afactorlab.com/