Al-Augmented Research teaches graduate students to integrate Al tools into scientific research with the same rigor they apply to experimental design and statistical inference. The course treats the use of large language models and prompting as an engineering discipline. It teaches retrieval-augmented generation, Al-assisted software development, agentic systems, human-Al collaboration patterns, and evaluation frameworks for detecting and analyzing Al failures. Students build working systems on their own research data throughout the course, culminating in a capstone project that produces a deliverable for their research group. The emphasis is on verification, failure analysis, and developing the professional judgment to know when Al helps, when it hurts, and when to set it aside entirely.
3 units · Letter or Credit/No Credit
Al-Augmented Research teaches graduate students to integrate Al tools into scientific research with the same rigor they apply to experimental design and statistical inference. The course treats the use of large language models and prompting as an engineering discipline. It teaches retrieval-augmented generation, Al-assisted software development, agentic systems, human-Al collaboration patterns, and evaluation frameworks for detecting and analyzing Al failures. Students build working systems on their own research data throughout the course, culminating in a capstone project that produces a deliverable for their research group. The emphasis is on verification, failure analysis, and developing the professional judgment to know when Al helps, when it hurts, and when to set it aside entirely.
Offered in Autumn 2026 at Stanford University.