AI for Finance: Machine Learning & Deep Learning for Trading

Hands-on AI trading: features, backtests, risk, and a paper-trading bot.

What you'll learn
- Build an end-to-end AI trading pipeline: data, EDA, features, leakage-safe splits, walk-forward tests.
- Train and tune ML/DL models (XGBoost, LSTM, Transformers) for forecasting and regime detection.
- Backtest event/news, sentiment, trend, and pairs strategies with costs, slippage, and risk metrics.
- Deploy a paper-trading bot with position sizing, volatility targeting, stops, monitoring, and ethics.

Requirements
- Curiosity—we’ll handle installs, tools, and paper-trading setup; no prior ML or trading required.

Description
Turn market ideas into working AI trading systems. In this hands-on course you’ll build a full pipeline in Python—from pulling real market and macro data to engineering features, training ML/DL models, validating with leakage-safe, walk-forward tests, and backtesting with realistic costs, slippage, and risk controls. You’ll implement multiple strategies (event/earnings & news, sentiment/NLP, trend/momentum, and pairs/stat-arb), compare models like XGBoost, Random Forests, LSTMs, and Transformers, and deploy a paper-trading bot with position sizing, volatility targeting, and clear monitoring dashboards. We work step-by-step in VS Code/Jupyter using pandas, scikit-learn, PyTorch, yfinance, vectorbt/Backtrader, and matplotlib—providing reusable notebooks, templates, and checklists so you can adapt everything to your own tickers and ideas. By the end, you’ll have reproducible workflow, a portfolio-ready project, and the confidence to iterate ethically and safely before going live.

Expect practical extras: a capstone project with template repo, model explainability (feature importance and SHAP-style reasoning), error analysis checklists, and hyperparameter tuning playbooks. We’ll cover data sourcing trade-offs, free alternatives to paid feeds, and pitfalls like survivorship bias. You’ll practice version control, experiment tracking, and reproducible runs, then stress-test results with regime changes. Optional extensions include crypto, options, and portfolio optimization. Support includes code reviews, troubleshooting tips, and a community.

(Educational use only—no performance guarantees.)

Who this course is for:
- This course is for curious self-starters who want to turn market ideas into working code—retail traders going systematic, developers/data analysts seeking a finance use case, students and career-switchers building a portfolio project, and fintech pros wanting hands-on ML. If you like learning by building—pulling real data, training models, backtesting, and paper-trading—this course fits. No prior ML or trading experience required; we guide you step-by-step.
More Info