firedup

I’m a quant who taught ML in finance at NYU for 4 years, and have spent ~6 months building an AI workflow that produces long-form equity primer reports. Not selling anything; I’m looking for blunt feedback from analysts who do equity research for a living.

What it does (brief)

    Pulls public sources (filings, transcripts, ownership/insider, patent/R&D mentions, sentiment).

    Outputs a structured primer (business model, KPIs, financials/ratios, risks, comp set, pipeline/innovation, catalysts).

    All points are citation-backed to public links; no paywalled sell-side.

Where I need your critique

    Biggest gaps vs. credible sell-side/independent work?

    Sections you’d cut/condense for real-world usability?

    Where would you not trust automation without a human pass?

    Preferred output format: single PDF/HTML vs modular notes?

Sample (mods: if links aren’t allowed, I’ll remove): RIG (Transocean) primer (HTML): https://storage.googleapis.com/derek-snow-at-outlook-co-nz-p...

Happy to answer technical/process questions in the thread. No DMs, no waitlists, just trying to make this genuinely useful for practitioners.

News Sentiment: Ticker-matched and theme-matched news sentiment datasets.

Price Breakout: Daily predictions for price breakouts of U.S. equities.

Insider Flow Prediction: Features insider trading metrics for machine learning models.

Institutional Trading: Insights into institutional investments and strategies.

Lobbying Data: Ticker-matched corporate lobbying data.

Short Selling: Short-selling datasets for risk analysis.

Wikipedia Views: Daily views and trends of large firms on Wikipedia.

Pharma Clinical Trials: Clinical trial data with success predictions.

Factor Signals: Traditional and alternative financial factors for modeling.

Financial Ratios: 80+ ratios from financial statements and market data.

Government Contracts: Data on contracts awarded to publicly traded companies.

Corporate Risks: Bankruptcy predictions for U.S. publicly traded stocks.

Global Risks: Daily updates on global risk perceptions.

CFPB Complaints: Consumer financial complaints data linked to tickers.

Risk Indicators: Corporate risk scores derived from events.

Traffic Agencies: Government website traffic data.

Earnings Surprise: Earnings announcements and estimates leading up to announcements.

Bankruptcy: Predictions for Chapter 7 and Chapter 11 bankruptcies in U.S. stocks.

How do you source them, google?

That is awesome, I have proposal for you, I currently run a niche notion website ml-quant.com, it could be much bigger and better than it is, the quant finance community is humongous and there is no central source for them, I have around 200 people on the website per day running all the important scrapers. Again this could be much bigger! I have a community both on a newsletter and on Linkedin. Let me know if this is something you would like to help me build, I would like to create a membership tier in 3-4 months. I am happy to go 50/50, you can get me on github for contacts https://github.com/firmai.

Interesting, unrelated, but related to your intro, what is the best open source datasets for maritime data?

The framework includes transformation from tensors, matrices, and vectors. It includes a range of encodings and decompositions such as Gramian Angular Encoding, Recurrence Plot, Markov Transition Fields, Matrix Product State, CANDECOMP, and Tucker Decompositions.

Github - https://github.com/firmai/datagene Colab - https://colab.research.google.com/drive/1QSDTKvNiwc1IRCX_VYr...