You are competing with a postdoc, not another offer
The standard assumption in hiring is that your candidate holds competing offers from similar companies. In quantum, the realistic alternative is usually another two or three years in a research group. That changes what your advert has to do: it is not arguing that you are the better employer, it is arguing that leaving research is not a mistake.
Candidates making that decision tend to weigh the same four things, roughly in this order: whether the technical work is real or a demonstration, whether they can keep publishing, whether the group around them is strong enough to learn from, and whether the move closes off a route back. Salary matters, but it rarely decides — someone who has spent a decade on a doctorate and two postdocs has already demonstrated they optimise for something else.
Their calendar is not your calendar
Academic careers run on cycles that have nothing to do with your quarter. Thesis submission, viva dates, grant end dates and fixed-term contract expiries all set the moment when someone can realistically move, and it is frequently months out. A candidate saying "not right now" is often saying "not until September", which is a very different answer.
-
1
Ask when they can actually start, early
Not as a filter but as information. A strong candidate available in six months is worth more than a weak one available immediately, and knowing lets you sequence the search rather than lose them.
-
2
Keep warm candidates warm deliberately
A short note when something ships, or an invitation to a talk, costs nothing and is how these hires usually land. The field is small enough that this is normal behaviour rather than pestering.
-
3
Expect a longer notice period than you are used to
Teaching commitments and grant obligations are real, and a candidate who honours them is showing you exactly how they will treat your commitments later.
Interviewing someone with no industry record
The usual signals are absent: no shipped products, no production incidents, no previous employers to reference. What there is instead is several years of documented, defended, independent work — which is a far better assessment surface than most industry CVs offer, provided you use it.
The single most effective interview for a research candidate is to ask them to teach you the hardest part of their thesis, at a level you can follow. It tests depth, honesty about limitations, and the ability to explain technical work to someone who does not share their background — which is, in practice, most of what the job will require of them. It also cannot be prepared for in the way a standard interview loop can.
What to assess, and how
Do not use a standard software interview loop on a hardware or algorithms candidate. Testing an experimental physicist on data structures tells you only that they have not practised for that test.
What converts, in practice
The offers that get accepted out of academia tend to share the same features, and none of them are unusual or expensive. They name a specific problem rather than a mission, they identify who the candidate will learn from, and they are explicit about publication and patents in writing rather than reassuring about them in conversation.
Name the problem
A concrete open problem attracts researchers in a way that a description of your funding round does not. They chose this field for the problems.
Put the research terms in writing
Publication rights, conference budget and patent policy, stated in the offer rather than promised verbally. This is the point at which offers are most often lost.
Before any of this, though, the role has to be scoped in a way that a single person can fill — which is where most quantum searches actually go wrong. How to hire quantum engineers covers the three tracks and how to choose between them.
See the market first
It is worth knowing what else a candidate is being shown before you write your own advert.