Conviction started as a question I kept running into as a student trying to research individual companies: the raw signals — who’s buying, who’s selling, what the fundamentals actually look like — are all public, but they live in completely different places, in completely different formats, updated on completely different schedules. Nobody outside a hedge fund had synthesized them into one thing you could actually look at.
So I built one.
The platform pulled together insider transactions (Form 4 filings — when executives and directors buy or sell their own company’s stock), institutional filings (13F filings — what hedge funds and large asset managers hold, updated quarterly), and standard company fundamentals, and combined them into a single Conviction score: a 0–100 read on how much conviction the signals collectively expressed about a company, with a transparent factor breakdown so the score was never a black box.
Around that core there was a full research terminal: a dashboard of market movers and top-scored names, per-company deep dives with a two-stage DCF model, a screener for filtering the universe by valuation and fundamentals, and dedicated views for insider activity and institutional positioning.
It ran for over a year as a real, deployed product — its own backend, database, and data pipelines — before I parked it to focus on other things. The screens below are from that build.