FlexStack: Python AI Principles 2 - Applied Generative AI

  • Overview
  • Course Content
  • Requirements & Materials
Overview

FlexStack: Python AI Principles 2 - Applied Generative AI

Course Description

As the second course in the FlexStack: Python AI Fundamentals Certificate, this offering teaches participants to design, build, and evaluate two working generative AI (GenAI) workflows. This includes retrieval-augmented generation for grounded answers with citations, and schema-validated extraction and triage for structured outputs. Learners will compare embeddings, tune chunking and indexing, implement abstain/refusal behaviors, and combine few-shot prompting with lightweight models. Labs add an evaluation harness with slice analysis and regression tests. They also cover monitoring, cost/latency tracking, and packaging to scripts with a runbook.

Course Content

WORKFLOW DESIGN 

  • Outcome-Driven Prompting and Gold Test Sets
  • Embedding Selection, Chunking, and Indexing
  • Retrieval-Augmented Generation (RAG) Pipeline with Citations and Abstain Rules 

STRUCTURED OUTPUTS 

  • Schema-Validated Extraction with Fallbacks
  • Triage Pipelines Combining Few-Shot, Weak Supervision, and Small Models 

EVALUATION AND OPERATIONS 

  • Evaluation Harness with Slice Analysis and Regression Tests
  • Monitoring of Cost, Latency, Error Buckets, and Drift
Requirements & Materials

Requirements

A Windows or Mac laptop with a webcam is required to participate in the course. Tablets or other devices are not supported. Recommended: Additional monitor. 

Familiarity with using a computer and adequate typing ability. 

Course registration requires an approved application and advisor meeting. 

Participants are expected to have their cameras on during the interactive sessions and must attend 80% of the sessions to pass the course.

Prerequisites

Required:

Successful completion of FlexStack: Python AI Principles 1 - Large Language Models.

Materials

Provided:

  • Canvas
  • Vocareum
  • H5P
  • Lecture Materials

Who Should Attend

This course is designed for technical professionals comfortable with Python. It is ideal for data and operations analysts, product and program managers, educators, Information Technology (IT) generalists, and technical leads who want to learn how to select and deploy practical, auditable solutions

Man who is a technical lead stands in front of screen with Python code as part of the FlexStack: Python AI Principles Certificate.

What You Will Learn

  • Outcome-driven prompting and test set design
  • Embedding selection, chunking, and indexing
  • Retrieval-Augmented Generation (RAG) workflows with citations and abstain rules
  • Schema-validated extraction and triage pipelines
  • Evaluation harness with slice and regression tests
  • Monitoring of cost, latency, and error buckets
  • Packaging to scripts with configs and runbooks
Woman sits in front of three monitors reviewing Python code with her male coworker for the FlexStack in Python AI Principles Certificate.

How You Will Benefit

  • Design outcome-driven prompts and gold test sets aligned to real tasks.
  • Compare embedding models and tune chunking and indexing to improve retrieval quality.
  • Build a baseline Retrieval-Augmented Generation (RAG) pipeline with citations and explicit refusal or abstain behavior.
  • Implement schema-validated extraction with fallbacks and confidence reporting.
  • Create an evaluation harness with slice analysis, regression tests, and release gates.
  • Monitor cost, latency, error buckets, and drift with clear alert thresholds.
  • Package workflows into scripts or a command line interface (CLI) with configs and a minimal runbook.

Want to learn more about this course?