Trading education roadmap

Algorithmic trading is automated trading: a script watches prices, applies rules, and sends orders without a human clicking buy or sell. A GitHub roadmap that spread this month lays out that path in Python, from software engineering and data science to finance and cloud deployment. (github.com) The repository, run by GitHub user rmcmillan34, has 148 stars and 18 forks as of April 12, 2026, and describes itself as a “self paced roadmap” for building trading systems powered by artificial intelligence and data science. Its top-level sections cover artificial intelligence, algorithmic trading, cloud and DevOps, and other supporting skills. (github.com) That roadmap matches how trading code is usually built in practice: first get historical market data, then test rules on old prices in a backtest, then run the same logic in a paper account before risking cash. Python libraries such as PyBroker package those steps into one workflow, including walk-forward testing that rechecks a strategy on fresh data instead of one fixed sample. (pybroker.com) Separate posts circulating on X pointed to a newer twist: large language models are being used not just to write code, but to read market questions, search the web, and decide whether to place a trade. Anthropic’s Claude documentation says its web search tool can pull in current web information with citations, which makes that setup possible. (platform.claude.com) One published example shows the mechanics in plain terms. Robot Traders posted a March 22, 2026 tutorial for a roughly 180-line Python script that takes a Polymarket market URL, asks Claude to analyze the question, and then trades or skips the market based on the model’s confidence. (robottraders.io) Polymarket’s own developer documentation shows why these experiments are easy to wire together: the platform exposes application programming interfaces for market data and trading on its central limit order book. Its terms of use bar disruptive practices such as spoofing, front-running, fictitious transactions, cornering, and wash trading. (docs.polymarket.com) (polymarket.com) Claims about returns have been much harder to verify than the code. BeInCrypto reported in March 2026 that some traders using Claude-based bots on Polymarket said the systems made profits ranging from thousands to millions of dollars during volatile political and macroeconomic events. (beincrypto.com) Those figures sit alongside a simpler fact: most of the reusable knowledge in these posts is about process, not guaranteed profit. The roadmap breaks trading into learnable modules, and the bot tutorials show how current tools can connect Python, live web research, and exchange application programming interfaces into one automated loop. (github.com) (robottraders.io)