🎙️ Podcast Link 🎙️

All #PhD recruitment approaches vary by professor, but there are some common themes that are fairly typical throughout.

In my newest #HackingAcademia video I shed some light on the #PhD candidate shortlisting and interview process we currently use, in the hope to demystify it for potential PhD candidates.

Some key points I cover include:

⭐ that what we’re looking for is just as much about fit and match to research in our group as it is to the particular individual involved, and we’re always appreciative of the time candidates put in

⭐ the huge number of applications we (and any lab with an international profile) get per year, and the practical necessities that forces upon us to be able to effectively manage them and shortlist – we do look at and consider just about every single applicant

⭐ what CVs etc… look like to us at a 30,000 ft view, and the irony that increasing professionalism means it’s harder to distinguish the best fit candidates from CVs alone (not that you should ever do that anyway!)

⭐ our two stage interview process for shortlisted candidates (including the occasional referee check, not mentioned in the video), consisting of

⭐ interview 1: “getting to know you” – tell us a bit about your background, what drew you to considering a PhD, what motivated you to apply to our lab specifically, what you understand the PhD process to be, and any questions you have about PhD life, and then

⭐ interview 2: technical knowledge and understanding, covering aspects of coding, computer vision, data structures, mathematics, combined with a “first year of your PhD crammed into 20 mins” mini-research study (in our topic area of Visual Place Recognition)

⭐ in the tech interview, you’re not expected to know everything, we provide extra support and scaffolding as needed (our aim is not to stress you out).

⭐ we’re particularly interested in you communicating what you’re thinking as you work through some of the questions and challenges, even if you don’t get it out at the end – in part, because this mimicks the whole student-supervisor dynamic in a PhD, which is so important

⭐ practicalities: interview processes are just as much about risk mitigation (in both directions) – not just about finding the absolute best candidate, although that’s also the aim – just not the sole one.

⭐ practicalities: we know good talent will likely apply to lots of labs, and have multiple options

Thanks especially to Tobias Fischer and Janet Danaher – we’ve tag teamed developing and refining the whole shortlisting and interview process, and will continue to tinker and improve!

Apologies for the jerky selfie cam – I’d been meaning to shoot this for months but there’s no time like the present.

#academic #research #jobs #phds #graduate #hiring #jobinterview #careers #careeradvice #university #undergraduate #HDR

Full Video Notes

One of the key activities an academic at a university goes through regularly is looking for and recruiting potential graduate PhD students. In many cases, a significant chunk—or even the majority—of research is carried out by PhD students, and pursuing a PhD is a rite of passage for many who aim for research, academic, or even industry research careers.

Every academic has a slightly different process, but the reason I’m making this video is to shed light on the particular process that our group currently uses. This process may change in the future, but I hope it provides useful insight for the tens, hundreds, or thousands of people considering a PhD each year.

First up, if you’re in a reasonably high-profile lab, you will likely receive an overwhelming number of PhD applications. Most labs with a significant international presence receive hundreds, if not thousands, of applications per year. While this is a fortunate position to be in, it means we need to be as efficient as possible in reviewing, tracking, and narrowing down applications to identify the students who would be the best fit for our lab—and vice versa.

These applications come in both solicited and unsolicited formats, typically including an email from the potential student, a CV, and sometimes a cover letter or expression of interest. In fields like robotics, computer vision, and machine learning, CVs tend to follow a standard format, listing educational details, research projects, internships, industry experience, papers, awards, and developmental activities like online courses.

The increasing professionalism of academia has made it challenging to assess, just from a CV, which applicants are likely to be a good fit for the research we hope to carry out. To address this, we shortlist candidates based on their CVs and other factors. At the moment, about one in 20 applicants makes it to the shortlist.

What do we look for when shortlisting candidates? A few things stand out. Ideally, candidates may have deep experience in the specific research field we’re working on—for example, visual place recognition. However, this is entirely optional, as it’s a niche area, and we don’t expect most applicants to have direct experience. Alternatively, we look for candidates with strong related skills, such as experience in computer vision, a solid mathematical background, exceptional software engineering skills, or expertise in machine learning areas relevant to our work. These related skills often indicate that the candidate could quickly get up to speed.

In our system, PhDs are relatively short—often three years flat. This structure means students need to hit the ground running, as there isn’t the luxury of a long, structured program with extended coursework.

Once we’ve shortlisted candidates, we reach out to set up one or two interviews. The first interview is a “get to know you” session, where we ask questions like why the candidate is interested in pursuing a PhD, their background and experience, and why they reached out to our group. While advice often recommends tailoring applications to specific labs and PIs, we understand that if you’re applying to hundreds of labs, this isn’t always feasible.

In this interview, we look for evidence that the candidate understands what a PhD involves. We also assess their communication skills, both verbal and written, which are critical for successful research. The candidate has the opportunity to ask questions about life in the lab, logistics like scholarships, and potential start dates.

After this interview, we confer and either thank the candidate for their time or invite them to a second, more technical interview. While we are genuinely grateful for the time candidates spend with us, the process inevitably involves narrowing down the pool further, recognizing that many strong candidates may choose other labs.

The technical interview has two components. First, we evaluate the candidate’s depth of knowledge and intuitive understanding of key technical concepts through questions about mathematics, coding, and machine learning. We don’t expect perfect answers but are interested in how candidates approach problems, think through solutions, and communicate their reasoning. We aim to make this part of the interview as supportive and constructive as possible, providing scaffolding and hints if a candidate gets stuck.

The second component is a mock research study. In this exercise, we condense the process of formulating, planning, and initiating a first-year research project into a 20-minute discussion. We explore how candidates think about acquiring datasets, evaluating system performance, and designing systems—particularly in areas like visual place recognition. There’s no single correct approach; the goal is to see how they think and how they might operate as a researcher. This part of the interview is particularly enjoyable because it mirrors working with our current students.

Following this interview, we either extend an offer to apply for a PhD or inform the candidate that we don’t have a suitable opportunity. It’s important to emphasize that this process is as much about finding a good fit in both directions as it is about evaluating a candidate’s capabilities. A lack of an offer often reflects the specifics of the lab’s needs at the time rather than the individual’s potential. We’ve seen excellent candidates go on to have stellar careers in other labs.

Another important aspect of the process is risk mitigation. While we aim to recruit top students, we also want to minimize the risk of mismatches that could lead to a poor experience for both the student and the lab. This isn’t about eliminating all risk, as that’s impossible, but about reducing the chances of incompatibility or misaligned expectations.

Different academics and labs have varied approaches to PhD recruitment. Some prioritize identifying raw potential and invest heavily in mentoring. Others, including us, aim to balance finding great candidates with mitigating potential challenges. If you’re applying to a specific professor, it’s worth tailoring your application to their stated preferences, but otherwise, don’t stress too much about trying to meet generic advice.

Our process is constantly evolving as we refine the questions and criteria we use. Ultimately, we’re grateful for the time and effort candidates invest in applying to our group, and we strive to make the process constructive and fair. Best of luck on your PhD journey!