From walkthroughs, to tutorials, to deep dives into AI orchestration, you're in the right place to learn more about how DAAF really works.
AI-empowered research is complex and multi-faceted, and we want to make sure you're equipped with everything you need to know to make the most of it and confidently stand by your work. Below, you'll find comprehensive guides to make the most of using DAAF: from first concepts to advanced extension. New to DAAF? After getting set up, we recommend reading these in order: Understanding DAAF → Best Practices → Extending DAAF → Philosophy & Vision.
Haven't installed DAAF yet? Our setup guide walks you through everything from installation to daily workflows. It only takes a few minutes to get up and running; make sure you start there before diving into the follow-up materials below.
How DAAF really works under the hood: context windows, engagement modes, the orchestrator-agent-skill architecture, and what a completed analysis looks like.
Read the guideWriting effective prompts, reviewing plans before execution, interpreting validation checkpoints, and understanding your oversight responsibilities as the researcher.
Read the guideHow to add your own datasets to DAAF, create new skills and agents, customize the Python and R environments, build custom workflows, and test your extensions end-to-end.
Read the guideBig-picture questions and discussion. Covers topics like the goals behind DAAF, the appropriate level of trust for AI-generated analysis, implications for equity and the next generation of researchers, the environmental costs of this work, and more.
Read the guideAn interactive, deep-dive walkthrough of a real end-to-end analysis with DAAF -- every step, every artifact, fully transparent. Well worth a read-through to build your intuition for what's going on under the hood and how it all fits together.
Explore the walkthroughNew to DAAF? Start with "Claude Code for Social Scientists: Get Started with DAAF in 30 Minutes" -- a full install walkthrough and orientation to the whole environment. Plus more step-by-step guides, explainers, and deep dives on YouTube.
Watch on YouTubeAn interactive benchmark explorer testing how well today's AI models actually operate DAAF -- routing requests, following protocol, delegating to subagents, and grounding their methodological advice. This is the place to check before experimenting with cheaper or open-weight models.
Explore the resultsThe DAAF Field Guide is a free Substack blog/newsletter dedicated to helping data scientists, data analysts, and researchers make the most of AI to deepen and extend their expertise and impact -- while enhancing the transparency, rigor, and reproducibility demanded by our core scientific principles.
Written by Brian Heseung Kim, DAAF's creator, the Field Guide offers approachable video tutorials, blog posts, guides, and deep dives that demystify complex AI concepts through accessible, intuitive explanations. Topics include what LLMs can reliably accomplish, how to communicate effectively with AI systems, coordinating multiple AI agents, managing context, understanding AI limitations, and agent orchestration.
Free for all readers and always will be!
What we can learn from high school English teachers and frontier AI model benchmarking
Read on Substack June 16, 2026Introducing DAAFBench: Orchestration!
Read on Substack April 20, 2026Do you really know what your AI agents are doing right now?
Read on Substack