TRDT

Sentinel: A Dual-Gated Alpha Engine. We merge Keras-based Neural Impact scoring with heuristic Advocacy logic to trade Polymarket. By requiring consensus between Math and Words, we cut the noise.

Comment
  • Wanting to explore the interconnection between journalism and trading markets. Quantitative data is one of the most common ways to analyse and identify market trends, so we want to quantify the qualitative aspects and see how strongly they can affect the market.

TRDT is a desk where you pick a live Polymarket event, see Google News stories ranked to that event, and run them through a signal pipeline to get a clear bullish/bearish-style verdict with explanations.

We built TRDT as a Next.js, TypeScript, Tailwind web app with server routes in front of Polymarket Gamma and Google News RSS, client-side ranking, bundling, and a signal pipeline with charts, plus a Python side that talks to Gamma, Data API, CLOB, and RSS for CLI checks and matching.

We fought slow, huge Polymarket payloads, finicky RSS, and messy news

We’re proud we shipped a live, event-first desk that wires Polymarket Gamma and news RSS into a clear, explainable verdict pipeline and backed it with a Python toolchain for the same APIs and matching.

Built With

Share this project:

Updates