Alexander Vershinin
Research pipelines, trading tools and AI agents, built in Python, JavaScript and React.
Flagship project
2025 – present. Hover over the chart, or tap it, to read each bar.
Python · machine learning · APIs
A research pipeline that cleans and analyzes 5-minute NQ futures data and tests breakout-based signals. Logistic Regression, Random Forest and neuroevolution models are evaluated out of sample on drawdown, win rate and risk-adjusted return (backtested). An API-integrated execution layer handles position sizing, logging and risk controls.
Claude · Python
Claude-based agents for market research, strategy ideation and code-generation support, used to speed up the research loop behind the trading work.
Builds
Trading apps, models and dashboards.
A trading app connected to Polymarket, with a FastAPI backend and a JavaScript front end. It runs on a Vultr VPS as a systemd service, so it stays up and restarts on its own.
A bot that analyses gold market conditions and outputs a risk level, from low to high.
A trading bot built in React that scans price charts for chart patterns.
A React dashboard with a dark terminal look that tracks tasks, Instagram analytics, gym training and a trading logbook in one place.
Markets
What I trade, how I study it and what I build with.
This page describes personal research and projects. Nothing here is investment advice or an offer to buy or sell any financial instrument, and charts use illustrative data. See the terms of use.
Contact