About

David Buckley is a UK-based software and AI engineer, transitioning from a career in physical education. This site documents a structured self-directed programme designed to develop production-quality engineering capability across the full modern AI stack.

The programme is organised into six phases covering CS fundamentals, the TypeScript web stack, applied machine learning, and deeper AI engineering including retrieval-augmented generation, agents, and MLOps. Each phase ships concrete artefacts: deployed applications, public repositories, and technical writeups. Progress against the programme is visible on the phases page and is computed at build time from the completion state of each phase's task list.

The curriculum draws on open, industry-respected resources including Harvard CS50, fast.ai, Andrew Ng's Machine Learning Specialization, Karpathy's Zero to Hero course, and the official Next.js and Vercel AI SDK documentation. All work is delivered in public and is available for technical review.

The objective is to demonstrate engineering capability through shipped work, with every artefact reviewable at source.

The scaffolding layer of this site was produced via AI pair programming with Claude Opus, driven through the Claude Code CLI. Programme artefacts shipped in subsequent phases — Python code, machine learning projects, deployed applications, technical writeups — are authored directly, with Claude used only as a code reviewer, tutor, and debugger. This distinction is explicit throughout the portfolio.