Who this is for
- R users who can write dplyr and want to build real ML and LLM systems. Comfort with the tidyverse is assumed; no prior ML or LLM experience needed.
Not for you if
- You've never written R — this assumes working tidyverse comfort
- You want a Python course — this is R, end to end
- You want theory and math derivations — this is applied and production-focused
- You want a five-minute prompting tutorial — this builds ML foundations first, then goes deep on LLMs
What you leave with
- A full tidymodels ML pipeline, trained, tuned, and evaluated
- A production RAG pipeline with proper retrieval and evaluation
- A deployed model API and a working Shiny + LLM chatbot
- The judgment to choose models, evaluate them honestly, and ship
The full curriculum
21 modules · 99 lessons · 4 phases. Every module below is listed with its lesson count; click any module to see its lessons.
Phase I · Machine Learning with tidymodels
From Visualization to Production ML
Build a rigorous ML foundation using tidymodels — from exploratory analysis through deployment. Sports data throughout. Real, production-grade workflows.
01Foundations of AI in R2 lessons
1.1ML vs LLMs
1.2AI Ecosystem in R
02From Viz to Models3 lessons
2.1Play-by-Play Data
2.2Exploration via Visualization
2.3Logistic Regression
03Model Evaluation and Testing6 lessons
3.1Training and Test Sets
3.2Model Specifications and Workflows
3.3Making Predictions
3.4Confusion Matrix and Basic Metrics
3.5Advanced Evaluation Techniques
3.6Overfitting and Test Evaluation
04Resampling and Model Tuning6 lessons
4.1Introduction to Cross-Validation
4.2Working with Resampling Results
4.3Comparing Models with Cross-Validation
4.4Introduction to Hyperparameter Tuning
4.5Selecting and Finalizing Models
4.6Tuning Multiple Parameters
05Predicting Continuous Outcomes4 lessons
5.1From Classification to Regression
5.2Linear Regression in Tidymodels
5.3Random Forests for Regression
5.4Evaluating Regression Models
06Preprocessing with Recipes4 lessons
6.1Why Preprocessing Matters
6.2Essential Recipe Steps
6.3Preparing and Applying Recipes
6.4Recipes in Workflows
07Regularized Regressions4 lessons
7.1The Problem of Overfitting
7.2Ridge, Lasso, and Elastic Net
7.3Tuning Regularized Models
7.4Interpreting Regularized Coefficients
08Workflow Sets for Model Comparison4 lessons
8.1The Challenge of Fair Model Comparison
8.2Creating Workflow Sets
8.3Tuning Workflow Sets
8.4Selecting and Finalizing the Best Workflow
09Neural Networks4 lessons
9.1Why Neural Networks
9.2Building Your First Neural Network
9.3Tuning Neural Networks
9.4Deeper Networks and Practical Considerations
10Deployment of ML Models4 lessons
10.1Why Deployment Matters
10.2Deploying Models with Vetiver
10.3Exporting Models with Orbital
10.4Monitoring and Maintaining Deployed Models
Career payoffA tidymodels + LLM production stack is exactly what R-heavy analytics teams are hiring for right now.
Phase II · LLM Foundations in R
Chat, Prompts, Structured Output, Multimodal
Build the core LLM skill stack in R using ellmer. Prompts, conversations, structured data extraction, multimodal input, and model selection including local models via Ollama.
11From ML to LLMs2 lessons
11.1A Gentle Introduction to LLMs
11.2Data Science Principles
12Chat Basics4 lessons
12.1Introduction to LLMs
12.2Getting Started with Ellmer
12.3Conversations, Roles, and Tokens
12.4System Prompts
13Prompt Engineering5 lessons
13.1Introduction to Prompt Engineering
13.2Adding Context and Instructions
13.3Handling Missing Data
13.4Few-shot Prompting
13.5Structuring and Storing Prompts
14Structured Output4 lessons
14.1Introduction to Structured Output
14.2Extracting Data Frames
14.3Handling Missing Fields
14.4Batch Processing
15Multimodal Input3 lessons
15.1Multimodal Input
15.2Hallucination with Plot Interpretation
15.3Extracting Data from Images
16Model Selection5 lessons
16.1Model Selection
16.2Working with Proprietary Models
16.3Local Models Intro
16.4Running Local Models with Ollama
16.5Local Models in R
Career payoffRAG and model evaluation move you from “uses AI” to “ships AI systems.”
Phase III · Production LLM Systems
Tool-calling, RAG, NLP, and Evaluation
The core differentiator of this course. Register R functions as tools. Build RAG systems with proper retrieval evaluation. Run NLP pipelines with mall. Measure everything with vitals.
17Tool Calling5 lessons
17.1Introduction to Tools
17.2Data Query Tools
17.3Data Query Tools (continued)
17.4Bio Retrieval Tools
17.5User Control and Safety
18NLP with Mall4 lessons
18.1Introduction to Mall
18.2Text Summarisation
18.3Extracting Information
18.4Text Classification
19Retrieval-Augmented Generation8 lessons
19.1Introducing RAG and Ragnar
19.2Document Processing
19.3Retrieval via BM25
19.4Retrieval via Embeddings
19.5Hybrid Retrieval
19.6Registering Retrieval Tools
19.7Context Augmentation
19.8Multiple Documents
20Model Evaluation with Vitals8 lessons
20.1Introduction to Evals
20.2Evaluation Data Sets
20.3Running First Evaluation
20.4Viewing Eval Results
20.5Comparing Models
20.6Statistical Comparison
20.7RAG Evaluation Design
20.8Running RAG Evaluation
Career payoffYou leave with deployed, evaluated apps — portfolio proof you can show an interviewer.
Phase IV · Shipping Real LLM Apps
Shiny + querychat + LLM Application Patterns
Put it all together. Build complete LLM applications with Shiny, querychat for NL-to-SQL, and shinychat for production chatbots. Tool-calling, RAG, model explanation — all wired into a real deployable app.
21LLM Apps10 lessons
21.1LLMs for Data Exploration
21.2Text to SQL
21.3Querychat
21.4Querychat System Prompt
21.5Customizing Querychat
21.6Chatbots with Shinychat
21.7Tool Calling in Shiny (Part 1)
21.8Tool Calling in Shiny (Part 2)
21.9RAG in Chatbots
21.10Explaining Model Predictions
What's included
- Lifetime accessEvery module and lesson, yours to keep and revisit, forever.
- Posit cloud workspaceA provisioned enterprise IDE and compute. Nothing to install.
- Project files and codeEvery notebook, dataset, and finished build, to keep and adapt.
- All future updatesNew lessons and refreshes as the tools move, at no extra cost.
- Self-pacedStart today, go at your own pace, no cohort to wait for.
- Free with MembershipOr get this course and the full catalog with Membership.
Your instructors
Nic Crane, PhD
Open Source Developer & R Educator | Apache Arrow Maintainer
Known for Apache Arrow Maintainer
Nic is an open source developer, R educator, and consultant. She maintains Apache Arrow, co-authored Scaling Up with Arrow and R, and builds courses teaching people how to work with LLMs in data science. Her work sits at the intersection of high-performance data tooling and practical, hands-on education for working data scientists.
LinkedIn →
Christoph Scheuch, PhD
Data Scientist & Product Manager | Creator of tidyfinance
Known for Creator of tidyfinance
Christoph is a financial economist and data scientist with deep, hypothesis-driven expertise in data wrangling, statistical analysis, and model development. He specializes in building scalable interactive dashboards (primarily in Shiny for R or Python, deployed with Docker, CI/CD, and ShinyProxy) and automated reporting pipelines built with Quarto and workflows like GitHub Actions. He builds custom open source packages including tidyfinance, fmpapi, and wbids in R and wbwdi and fmpapi in Python. With extensive teaching experience from university courses to corporate training, he delivers engaging, practical sessions across data visualization, data transformation, and reproducible workflows, adapting content and pace to academics, analysts, and business professionals alike.
LinkedIn →Questions
Is this included with Membership?
Yes. AthlyticZ Membership ($2,999/yr) includes this and the entire course catalog, plus the live Masterclass. If you plan to take more than a couple of courses, Membership is the better math.
Is it self-paced?
Yes. Start today and go at your own pace, with lifetime access to every lesson and all future updates.
Do I need to install anything?
No. You work on the provisioned Posit platform in the browser, on the same enterprise tools professional teams use.