What You'll Learn

• Build autonomous AI agents with the ReAct pattern: Understand how to create agents that alternate between reasoning about what needs to happen next and taking action with tools, eliminating manual function chaining.
• Design LangGraph workflows for machine learning pipelines: Use StateGraph to orchestrate sequential ML operations with conditional routing, tool nodes, and state management that scales beyond simple scripts.
• Create custom tools for data cleaning, feature selection, and model training: Learn to wrap ML operations in LangChain tools that language models can intelligently invoke, including handling data leakage, categorical encoding, and missing values.
• Implement probability based classification with XGBoost: Go beyond simple yes/no predictions to generate match probability scores, and evaluate model quality using ROC AUC metrics that measure ranking performance.
• Handle real world data challenges in speed dating datasets: Clean byte string encodings, remove data leakage from decision variables, impute missing values, and use Recursive Feature Elimination to identify what actually predicts romantic chemistry.

Who Should Enroll

• Data scientists and ML engineers who want to automate repetitive pipeline steps and understand how agentic AI can orchestrate complex workflows without constant manual intervention.
• Python developers interested in AI agents who want practical, production relevant experience with LangGraph and the ReAct pattern beyond simple chatbot demos.
• Students and professionals exploring machine learning automation who need to understand how modern AI systems can reason about multistep tasks and make intelligent decisions about tool usage.

Why Enroll

What You'll Need