Free ML/quant resources surfaced

Free machine-learning and quantitative-finance study materials are circulating together online, bundling beginner coursework with open-source trading code and strategy tutorials. (developers.google.com) (github.com) Machine learning is software that learns patterns from past data, and in markets that usually means fitting models to prices, volumes, or company data to forecast or classify moves. Google’s Machine Learning Crash Course is a free self-study program that covers regression, classification, data handling, neural networks, and overfitting. (developers.google.com) Google says the course was refreshed in November 2024 with new material on large language models, automated machine learning, responsible artificial intelligence, and more than 130 exercise questions. The lessons run in a browser through Colaboratory, and Google says no prior machine-learning knowledge is required, though Python and basic algebra help. (blog.google) (developers.google.com 1) (developers.google.com 2) The finance angle comes from projects that translate those basics into trading workflows. One GitHub project, AutoHedge, describes itself as an “autonomous agent hedge fund” with specialized agents for strategy, quantitative analysis, risk management, and execution. (github.com) As of this week, that repository showed about 1,200 stars, 219 forks, 37 commits, and support for autonomous trading on Solana, with Coinbase listed as “coming soon.” The project’s README says it is built around live market analysis, structured outputs, logging, and a “risk-first architecture.” (github.com) Another strand of the bundle is statistical arbitrage, a trading style that looks for pairs or baskets whose price gap tends to snap back like a stretched rubber band. One public Python example uses cointegration tests, z-scores, hedge-ratio calibration, and backtesting rules such as entering when a spread moves beyond two standard deviations. (github.com) That mix of free lessons and public code arrives as creators pitch do-it-yourself quant tooling to retail traders. Moon Dev’s YouTube channel, which posts frequent videos about Claude-built bots and automated trading systems, showed about 106,000 subscribers and more than 530 videos when checked on April 15, 2026. (youtube.com) The materials vary in rigor: Google’s course is an official training product, while GitHub repositories and creator tutorials are user-published and can change quickly. Even the better-documented examples focus on building and testing systems, not proving that any strategy will stay profitable in live markets. (developers.google.com) (github.com 1) (github.com 2) What surfaced this week was not a new academic breakthrough but a starter kit: learn regression and classification, run browser exercises, inspect open-source trading agents, and study how pairs-trading code turns statistics into signals. The through line is access — practical machine-learning finance material that costs nothing to open. (developers.google.com) (blog.google) (github.com)