Week 1 is open to everyone. It builds a working understanding of how large
language models and agents behave, then puts that understanding to work with
Claude, Claude Cowork, and Claude Code.
Objectives
By the end of the week you should be able to:
- Explain, at a working level, how LLMs and agents operate.
- Use Claude and Claude Cowork for research, reporting, brainstorming,
red-teaming, and admin tasks.
- Apply context-engineering strategies and use skills effectively.
- Use Claude Code for scientific programming, from the basics through to
running parallel instances with remote MCP.
- Use Claude Code's built-in skill for converting codebases across languages.
The week at a glance
- Days 1–2 (no prerequisites): how LLMs and agents work; using Claude and
Claude Cowork for research, reporting, brainstorming, red-teaming, and admin;
an introduction to skills and context engineering.
- Days 3–5 (scientific programming interest and ability): Claude Code, from
the basics to effectively running parallel instances with remote MCP — with a
highlight on converting codebases across languages.
How this feeds your project
The cowork skills from this week are how your team will research, synthesize,
and report throughout the course. The Claude Code skills are how you'll do the
actual scientific development for your project.