Start Here: The Path to AI Research
What AI researchers actually do, the realistic routes into a funded PhD, what selection panels look for, and a plan for your next 12–18 months.
What an AI researcher does
A researcher's job is to produce new, trustworthy knowledge and share it so others can build on it. In AI that usually means one of: a new method that works better or more cheaply, an analysis that explains why something works or fails, a new dataset or benchmark that measures something that matters, a theoretical result, or a careful application to an important problem (health, climate, education) that teaches the field something.
The day-to-day work is less glamorous than the headlines: reading a lot, thinking about what is actually unknown, running many experiments that fail, checking results until you trust them, and writing clearly. The skills that matter most are taste (choosing questions worth answering), rigour (not fooling yourself), persistence, and communication. All four can be learned, and this course is organised around them.
Routes into a funded PhD
Most funded AI PhD students got there through one or more of these:
- Strong grades plus a research project: an honours, master's or capstone thesis with a supervisor who can write a detailed reference about your research ability.
- A research assistant role (paid or volunteer) in a lab for 6–12 months, often leading to a co-authored paper.
- A first paper: a workshop paper, a short conference paper, or a strong arXiv preprint with code. Not required everywhere, but it is the clearest evidence that you can do research.
- Fit with a supervisor: in Australia, the UK and much of Europe you usually apply to work with a specific supervisor on a topic in their area; in the US you apply to a department, but naming faculty you fit with still matters.
For international applicants, competition for funded places is strong, so the evidence matters even more: a clear research record, specific fit, and references who know your research. The good news is that every piece of that evidence is something you can build deliberately, and that is what this course walks you through.
What selection panels look for
| Evidence | Why it matters | How you build it |
|---|---|---|
| Academic record | Most scholarships rank on grades (e.g. a GPA or "first-class honours equivalent") | Prioritise research-relevant units; explain anomalies briefly |
| Research outputs | Shows you can finish research | Thesis, workshop paper, preprint, code release (the capstone) |
| References | Panels trust specific, detailed letters most | Work closely with 2–3 academics; give them material |
| Proposal and fit | Shows you understand the field and the supervisor's work | Literature review + a focused, feasible proposal (Part 5) |
| Relevant experience | Industry, RA work, competitions, teaching | Describe what you did and learned, not just titles |
Many Australian universities rank scholarship applicants with a points system that weights grades and research outputs; read your target university's ranking guidelines, because they tell you exactly what counts.
Try it: plan backwards from your target intake
Choose when you want to start a PhD and how many hours a week you can give. The planner works backwards: capstone paper, supervisor contact, references, proposal and scholarship deadlines.
Interactive: this part runs in your browser.
How to use this course
- Parts 1–3 teach the craft. Each topic has a quiz; each part ends with a checkpoint.
- Start the capstone workbook early, alongside Part 2: choose your topic while you learn how to design experiments, and fill in each stage as you reach the matching topic.
- Play the game milestones as you reach them; they practise the decisions in realistic situations.
- Part 5 turns your capstone into a proposal, a statement of purpose, supervisor emails and scholarship applications.
Find a mentor if you can: a lecturer, a PhD student or a researcher who will read your drafts. This course can teach the process; feedback from a person who knows your field is still irreplaceable.
Quiz
1. Which is the clearest single piece of evidence that you can do research?
2. In Australia and the UK, PhD applications usually…
3. A reference letter is most valuable when it…
4. Where do you find exactly how a university ranks scholarship applicants?
📚 Further reading
Trusted, mostly free sources: official documentation, well-known textbooks and courses. When you want to confirm or go deeper on anything here, start with these.
- A. Karpathy, "A Survival Guide to a PhD" ↗ choosing problems, writing, and PhD life
- J. Eisner, Advice for Research Students ↗ a curated collection of advice on reading, writing and research
- P. Guo, The PhD Grind ↗ a free memoir of a PhD