🎙️ Podcast Link 🎙️
I’ve long maintained a webpage [1] that details what a #PhD is really like on a day to day basis in #robotics or related fields and in the Australian PhD system, but a lot of potential students we interact with like to hear us talk it through verbally.
So today, taking advantage of both the #WFH mandate and the much-delayed landfall of #CycloneAlfred, I shot a #HackingAcademia video in my garden. Or at least starting in my garden, then transitioning undercover when it started raining!
This informal video (the first 15 minutes is in this post, for full video go to YouTube) lays out:
💠 the milestones and timeline
💠 forming your research question and subquestions
💠 submitting and publishing academic research conference or journal papers
💠 working with supervisors, meeting regularly, and having more than one for redundancy and extra support and perspectives
💠 getting involved in the culture and activities of the research group and larger institution you’re within
💠 opportunities for conference and internship travel
💠 leveraging existing skillsets and experience where possible, given the relatively short nature of an Australian PhD
💠 the paradox that for many a PhD is simultaneously the largest project they’ve yet taken on, but also much narrower, and more focused, then they initially envisaged
💠 dealing with setbacks and support for Plan Bs especially from your supervisory team
💠 the “bursty” nature of a PhD, especially around publication deadlines
…and many other key aspects of the PhD process – focusing mainly on the process after you’ve commenced (other videos cover the application stage: see https://michaelmilford.com/hacking-academia-tutorial-series/).
Again, this is specific to the Australian context – most notably the relatively short duration compared to the United States for example – and to topic areas like robotics and related fields. It will also of course still vary from lab to lab.
The video thumbnail is from my PhD graduation, a looooooong time ago – and no, this video isn’t about my experience back then, but about the contemporary, current experience. That said, it’s interesting that not absolutely everything has changed radically about the process since then.
[1] https://michaelmilford.com/what-a-phd-in-robotics-is-like/
#PhD #graduate #studentlife #university #research #robotics #computervision #machinelearning #artificialintelligence #degree #learning #career #careeradvice #careerdevelopment #perception #manipulation #grasping #navigation #localization #SLAM #mapping #positioning #reinforcementlearning #robot #robots #autonomousvehicles #selfdriving #agriculture #construction #mining #medicine #health #upskilling
Full Video Notes
Since Cyclone Alfred has not yet hit Brisbane, I wanted to take the chance to get outside. It is currently a beautiful, sunny day—believe it or not—and I thought I’d tell you a little bit about what it’s like to do a robotics PhD in our lab. While this is based on our experience, much of it would also apply to robotics or similar discipline PhDs in other labs across Australia.
I’m not going to focus on the application or shortlisting process in this video. I’ve covered that in other videos. Instead, this one is about the actual experience of doing the PhD once you’ve started.
A typical PhD student comes to us through one of two broad pathways. The first is as a domestic student—someone already in Australia, either coming straight out of an undergraduate degree or returning from work or overseas study to do a PhD. The second is as an international student—someone coming from overseas, such as India, to undertake a PhD in Australia.
The Australian PhD program is relatively short by international standards. Unlike the United States, where PhDs often stretch to five or six years, in Australia, most students aim to finish in three years, or sometimes three and a half. While this doesn’t always happen, it is the nominal goal, and some students even finish earlier.
Because the program is short, there is generally not much formal coursework involved. You might take some basic training modules on things like information retrieval, but there is no large body of structured coursework. This compressed timeline has implications. One is that we try to select students who can hit the ground running. During the interview and selection process, we pay close attention to your existing skills and experience. Where possible, we shape the PhD topic to align with your background so that you can make progress from an earlier stage.
This doesn’t mean you won’t learn new things. A PhD is still a huge learning opportunity. But if you are starting entirely from scratch, it can be difficult to build up the necessary knowledge quickly enough in a three-year window. So when you have relevant skills, we aim to leverage them.
Once you begin the PhD, you will work closely with your supervisors. We always aim to assign at least two actively involved supervisors: a principal supervisor and an associate supervisor. Both should be capable of guiding your research, and this helps ensure continuity if one is unavailable due to travel or other commitments.
The PhD program has a series of formal milestones. First, there is an initial proposal that you submit when applying, where you outline your topic and plan. About three months into the program, you submit a more detailed research plan describing your overarching research question, subquestions, methodology, and background, as well as the resources and support you’ll need. At the one-year mark, you submit a major report and deliver a 45–50-minute presentation outlining your progress.
Toward the end of the program, you’ll submit your final thesis and give a concluding presentation. Your thesis can be structured in one of two ways. The first is a thesis by publication. In this format, you write an original introduction and background chapter, then include several of your published or submitted papers as chapters, and conclude with a general discussion chapter. The second option is the traditional monograph thesis, where you write the entire document from scratch. Once submitted, your thesis is reviewed by external examiners, who provide feedback and a recommendation. Most commonly, this includes a request for revisions, which you complete and resubmit before being awarded the PhD.
To explain more about the structure: your research will be guided by one overarching research question, broken into three subquestions. These often align roughly with each year of your PhD. The first subquestion is typically something you already begin investigating early on and have a good feel for. The second and third are more speculative, and your work on the first may shape or influence how you approach the others.
You will articulate these questions at various stages. You might start with a high-level version during your application, then expand it in detail at the three-month milestone, and by the one-year point, you should have a well-defined plan. Your research directions may shift during the process, especially as you start answering the first question and realize adjustments are needed. That is completely normal.
During your first year, you’ll meet regularly with your supervisors—typically weekly. Meetings may be in person or online, depending on schedules and travel. If one supervisor is unavailable, the associate supervisor can step in. When both are away, we make sure you have a clear plan to follow during their absence so that your progress doesn’t stall.
The first year is particularly important. Traditionally, the early stage of a PhD was mostly focused on reading and background work. However, we now try to get students working on actual research tasks early—writing code, building algorithms, working with robots or datasets—while reading in parallel. This dual focus helps you engage with your research area more deeply and avoids burnout from only reading papers.
As your PhD progresses and stabilizes, we encourage you to explore collaboration opportunities within the lab. One benefit of being in a large research centre is the diversity of people and projects around you. You may also take on part-time paid work, like tutoring or assisting with lab sessions. We generally recommend keeping this to a minimum in the first year, but it can be a good opportunity as you gain momentum.
Most students will submit their first paper around eight or nine months into the program, based on their early work related to the first subquestion. This paper would typically be submitted to a conference or journal. Of course, timelines vary. Some students submit their first paper after one and a half or even two years, depending on the project and circumstances.
You will likely attend at least one national or international conference during your PhD, where you’ll present your first-author work. These trips can be incredibly valuable for networking and exposure to global research. Depending on budget and progress, some students also do internships or research visits—often for three months with a collaborating lab overseas. These experiences can be life-changing, especially for students who have not spent much time outside their home country.
Back at your home lab or centre, there are many opportunities for professional and social engagement. We have regular journal clubs, discussion groups, training workshops, and even social events like board game nights or philosophy meetups. These activities are great for building community and are especially important for international students adjusting to a new environment.
The PhD journey rarely goes exactly as planned. A common example is submitting a paper and getting rejected from the first conference. This can be discouraging, especially the first time, but it happens to nearly everyone. In these moments, your supervisors are key. They can help you improve the work, revise it, and resubmit to a more suitable venue. Most of the time, these things resolve themselves with persistence.
Collaboration also becomes more prominent in later stages of the PhD. Early on, we try to limit side collaborations to keep your core research on track. But as things progress, collaborative work can add depth and value. Sometimes it will directly contribute to your thesis, other times it’s more about broadening your experience. Your contribution might be based on your particular skills, such as expertise with a robot platform, coding, mapping, or navigation.
People often ask about work-life balance during a PhD. While the overall workload might not be higher than a typical full-time job, the intensity is often more variable. Workload spikes tend to happen around deadlines, especially paper submissions, which can be stressful. However, quieter periods after a submission can offer time for reflection, idea exploration, and recovery.
From a broader perspective, a PhD is almost always narrower than students expect. I often tell students that the PhD will be both the biggest project they’ve ever done and yet surprisingly small and focused. Early on, we work closely with students to narrow their topic, cutting down the scope and setting stretch goals that are rarely reached. Many students look back proudly on what they’ve achieved but also wish they had time to explore more ideas. This is a normal part of the process.
You will learn a tremendous amount during your PhD. On the technical side, you will improve your coding, mathematical, communication, and writing skills dramatically. You will also develop valuable soft skills—time management, project planning, stakeholder engagement, and interpersonal communication. You may even deal with challenging professional relationships or stressful situations, all of which build resilience and maturity.
These skills are valuable in any career path, whether you continue in academia, move into industry, or join a startup. Your PhD experience will shape you professionally and personally.
That’s it for this informal overview of what a robotics PhD is like in our lab—and in many others across Australia. I hope it gives you a clearer sense of what to expect. I didn’t cover everything, and I didn’t go into much detail about the application process. But I hope this helps reduce some of the uncertainty around what it’s like to actually do a PhD in this environment.