Thou Shalt Learn by Building
Introduce advanced tools and concepts early so projects can start immediately, then revisit the same subject later for deeper understanding.
A teaching methodology shaped by patterns from startup engineering and industry R&D leadership: build first, formalize later. These 10 principles guide the design of the Hands-on AI Science courses.
Introduce advanced tools and concepts early so projects can start immediately, then revisit the same subject later for deeper understanding.
Teach theory through code examples, treating code as a first-class citizen tightly connected to concepts, with practice and theory fully interwoven.
Teach state-of-the-art models and tools rather than yesterday’s curricula or waiting for a consensus syllabus.
Encourage students to invent and build new things; innovation is mandatory, while standard tasks are delegated to AI.
Teach a broad, practical toolkit for solving real problems instead of deep-diving into only a few isolated concepts.
Students are encouraged to reuse components and delegate coding or writing to AI, while remaining fully responsible for every design decision and outcome.
Teach students to explain ideas and innovations clearly in words, so both people and AI can understand and act on them.
Provide continuous feedback to guide projects toward clearer problem definitions, better scope, and stronger implementations.
Ensure students leave with a demonstrable portfolio project, including a code repository and presentation artifacts.
Teaching works best when both instructor and students enjoy the process. Keep the energy high, the examples vivid, and the problems genuinely interesting.