FAQ

How long will this course take?

Plan on four to eight weeks end-to-end, depending on your background and time. Each unit is about three to five hours of focused work for Weeks 1 and 2, or five to seven hours for Weeks 3 and 4. Some learners compress a unit into a long weekend; others spread it over two weeks. There are no deadlines. Pick a pace you can sustain.

Do I need a cohort or instructor?

No. The course is designed to be taken solo. Every exercise has a self-check answer key, every project has a learner-facing self-rubric, and every conceptual page has a “Check your understanding” callout. See How to use this course for the self-assessment model.

If you’d prefer to take it with peers, it also works well as a small study group of 3 to 5 people who agree on a shared cadence. That’s optional, not required.

How do I know if I’ve understood the material?

You have four kinds of feedback:

1. Per-page checks at the bottom of every conceptual page (3 to 5 questions with worked answers).
2. Per-week knowledge checks at the end of each week page (5 to 8 mixed questions).
3. Hands-on practice exercises that apply each unit to your own work.
4. Self-rubrics for each project, so you can grade your own work against the same dimensions an instructor would.

If you miss two or more questions on a topic in the knowledge check, revisit the reading and redo the practice. The check is the signal. The reading is the fix.

Can I skip a week?

Each week is a unit, and they build on each other.

• Week 1 (the 4 D’s) is the framework everything else hangs off. Don’t skip it.
• Week 2 (LLM literacy) is foundational for the prompt-engineering and hands-on tracks. If you skip it, your prompts in Weeks 3 and 4 will be weaker.
• Weeks 3 and 4 are the hands-on bioinformatics track. If you are not in genomics and don’t care about the scRNA-seq pipeline, you can substitute a Week 4 final project from your own subfield. Path B is a literature brief, and Path C is protocol design (see Week 4).

What if I get stuck on the project?

For the Week 3 mini-project, a Python starter confirms the dataset loads on your machine, and Modules 3 and 4 are runnable Colab notebooks with prompts and gold-standard outputs. If your AI-free baseline isn’t running, that’s the signal to slow down. Debug the baseline before adding AI assistance, because the AI will paper over an unclear bug.

For all projects, the self-review template and the disclosure rubric in the syllabus give you concrete questions to ask of your own work. If you can answer them, the project is done. If you can’t, that’s where the work is.

Do I need a paid AI subscription?

No. The course is designed to work with free-tier access. Paid tiers (Claude Pro, ChatGPT Plus, and others) give you more usage headroom and access to more capable models, which makes the hands-on units smoother, but they are not required. Free-tier rate limits may interrupt sustained code-assistance sessions in Weeks 3 and 4. Plan accordingly.

Which AI tool should I use?

The course examples use Claude by default, because its behaviour and long-context handling suit research workflows well. Most exercises work with any frontier LLM. Where a tool-specific feature matters (Claude’s artifacts, tool use, or computer use), the page flags it.

Can I use AI for the projects?

Yes. That’s the whole point. See the AI-use policy in the syllabus for disclosure and verification expectations. Two projects (Week 1 and Week 3) include an AI-free baseline component, so you build the underlying skill before collaborating with an AI. That’s a pedagogical choice, not a “no AI” rule.

I’m not a bioinformatician. Will I keep up?

Yes. The conceptual tracks (fluency, literacy) work for any stack. For the hands-on track, the bioinformatics examples are scaffolded so the AI does most of the mechanical work and you bring the biological judgement. If you don’t have a genomics background at all, Week 4 lets you choose a non-genomics path (literature brief or protocol design) for the final project.

I’m an advanced bioinformatician. Will I be bored?

The mechanics might be familiar, but the framing (delegation, discernment, epistemics, disclosure practice) is what most working bioinformaticians are missing. Treat it as professional development. The “Check your understanding” callouts are a fast filter. If you breeze through them, skip the reading and try the practice exercise.

Is there a certificate?

No. The course is open materials, with no enrolment, no exam, and no certificate. If you want something to point to after finishing, your completed final project (with its disclosure statement and self-review) is the artifact. Put it on a GitHub repo and link it from your CV. That’s a stronger signal than a certificate.

What about data privacy?

Never paste confidential, embargoed, patient-identifiable, or unpublished-collaborator data into a third-party LLM without explicit permission and appropriate privacy safeguards. The course does not require you to upload any sensitive data. The PBMC 3k dataset is fully public. See the data policy in the syllabus for the full list.