Ask a quantum start-up what it cannot hire and the answer is rarely the one the brochures expect. In a study of 34 companies published on 31 July 2024, representatives of 12 said they faced a talent shortage, and one put the problem in a sentence most physics graduates would not predict: "for us it’s mostly hard not to get the quantum engineers, but to get to the software engineers." That sentence is one voice. It is also, as this piece shows, where every dated public document that ranks anything points.
This is the last piece of the quantum chapter of the CareerOn Industry Atlas. The first followed the money and found the French State acting as planner and buyer. The second followed the hardware underneath. The third read what Europe is actually building in 2026, and the fourth read the arguments inside the field. This one asks the question a candidate actually brings: which quantum jobs exist, which are scarce, and which of them a person can rehearse before the interview. We print what each document measures beside what it says, we add nothing across documents that measure different things, and we say plainly where CareerOn cannot yet help.
The quantum skills gap is not where the field's image puts it: the documents that rank shortages put software engineering, system integration and quantum cybersecurity first and find French training oriented to physics, while engineers for qubit control electronics are named more often than any other engineering profile; no admitted document counts the open roles, so the honest map is a ranking in words, one simulation CareerOn can place against named work, and five named jobs it cannot yet rehearse.
We read eight dated documents across four families. Two are French public documents read again from earlier pieces: the Cour des comptes audit of 20 July 2026 and the parliamentary mission report of January 2020. One is a French trade body report, from the aerospace and defence association GIFAS, of May 2024. Two are European public documents: the Commission's Quantum Europe Strategy of 2 July 2025, read in a real browser because the official site refuses scripted reads, and the European Competence Framework for Quantum Technologies, in its April 2024 edition. Two come from research and industry: an interview study from TU Braunschweig, dated 31 July 2024, and the recommendations of the European Quantum Industry Consortium of March 2025. The last is international: the OECD's map of the global quantum ecosystem of December 2025, which reads job postings in Canada, the United Kingdom and the United States.
We refused three documents. A widely quoted consultancy monitor did not open to our probe, and its talent figures compare postings with graduates, a ratio we will not restate as a shortage. A publisher's page for the interview study answered with a web page rather than the paper, so the study is read from its dated preprint. A summary page of the Commission strategy is a library entry; we read the Communication itself.
What the documents can and cannot count
The first thing to say about quantum jobs is the thing most articles skip: nobody we could read counts them. The Commission ranks shortages in words and prints no size. The audit judges a training offer. The interview study counts companies that mention a problem, which is not the same as a share of a labour market. The OECD comes closest to demand, and it is careful about what its data are: "The data span from January 2021 to July 2025 and capture job vacancies posted online during this period, reflecting positions advertised rather than the existing workforce stock." Those postings are in three countries outside the European Union. We do not print them as French or European demand.
The industry consortium makes the forecast the field likes to repeat: "The gap between the talent pool and demand from the job market is expected to widen considerably over the coming years." It is an industry body speaking for its members, and it measures nothing. The French documents are recommendations of the same kind. The 2020 mission asked the country to "anticiper la croissance du besoin en ingénieurs et techniciens des filières industrielles." GIFAS, in May 2024, asked for "une offre importante de formation continue dans le domaine des technologies quantiques à destination des techniciens et des ingénieurs ayant vocation à travailler dans le domaine." The mission report is quoted again by the audit, so the two are one line of French evidence, not two.
This matters for a candidate because a number would change the advice. If the open roles were counted, one could say how many software engineers the field needs against how many hardware engineers. Without that count, the strongest available statement is an order: which shortages the documents call most critical, and which profiles employers name most often. That is what we print, and we print it as an order, not as a size.
Where the documents put the gap
The Commission's strategy of 2 July 2025 is the one public document that ranks. It writes: "Shortages are most critical in applied fields 78 , including quantum software engineering, system integration, and quantum cybersecurity, slowing the commercialisation path for EU-based startups and scaleups." The number in that sentence is a footnote marker in the Communication, kept here because we quote verbatim. The three fields it names are not physics research. They are the work of turning a machine into something a customer can program, connect and trust.
The Cour des comptes, writing about France in July 2026, reaches the same place from the supply side. Its judgement on the training offer is that it "semble cependant très orientée sur la physique et moins sur le volet informatique." It adds that the needs "portent aussi sur les volets « architecture » et « correction d’erreur », ainsi que sur l’articulation entre informatique classique et informatique quantique." The audit also counts one supply line, from the doctoral programme funded with industry: 149 doctoral candidates trained between 2017 and 2025, 77 % of them French and 9 % from elsewhere in the European Union, mostly on quantum computing (29 %), software (24 %) and cryptography (23 %). That is a count of people trained over nine years, not of jobs open now, and we print it as such.
The interview study gives the ranking a voice. Its authors spoke to representatives of 34 companies, mostly in Germany and France, and 12 mentioned the shortage "or the difficulty of attracting talent, representing all types of companies." The software sentence quoted at the top of this piece comes from that study. Three documents, three kinds of evidence, one direction: the sharpest gap the documents describe is on the software and systems side of the machine.
The documents that rank anything put software and systems first
| What the document prints | Kind of evidence | Scope | Publisher, date |
|---|---|---|---|
| Shortages most critical in quantum software engineering, system integration and quantum cybersecurity | Policy ranking | EU | European Commission, July 2025 |
| Training very oriented to physics, while needs also cover architecture, error correction and the classical-quantum link | Audit finding | France | Cour des comptes, July 2026 |
| 12 of 34 companies mention a talent shortage; one says the hard hire is the software engineer | Interview count | Europe, interviews | TU Braunschweig study, July 2024 |
| Electronic engineers for qubit control named more often than any other engineering profile | Interview count, counter-case | Europe, interviews | TU Braunschweig study, July 2024 |
Sources 4 European Commission · 1 Cour des comptes · 6 Greinert, Ubben, Dogan, Hilfert-Rüppell and Müller (TU Braunschweig, QUCATS); arXiv 2407.21598, later EPJ Quantum Technology
The counter-case is in the same study, and we print it at full strength. When the authors list the engineering profiles companies want, "Electrical and electronic engineers, including electromagnetic wave engineers or engineers for micro- or analogue electronics, e.g. for qubit control, are mentioned more often than e.g. mechanical engineers, microwave engineers, optical engineers, or ‘other engineers’ like FPGA engineers, cryo or telecom engineers." Hardware control is in demand too. A reader who concludes that only software matters has misread the evidence. What the documents support is narrower: among the shortages anyone ranks, software and systems come first, and among engineering profiles employers name, control electronics comes first.
Which job you can rehearse today
CareerOn places a simulation against a topic only when two admitted documents name the work it rehearses, in their own words. Two quantum simulations exist in the catalogue today. One passes that test and one does not, and we print both results.
The competence framework describes an engineer who "works on the development or improvement of QT facets, on the hardware or software side, e.g. on improving the control electronics for a qubit realisation." The interview study names "engineers for micro- or analogue electronics, e.g. for qubit control." That is the work the cryogenic control simulation rehearses. Neither document names the cryogenic stage itself, so we say the simulation rehearses one way of doing that work, not the only one.
There is a disclosure that belongs beside that placement. Both documents come out of the same EU-funded qualification effort: the framework is published under the Quantum Flagship, and the study's authors work in the same programme. They are two documents from one programme, not two independent witnesses. We placed the simulation because both name the work verbatim; we tell the reader where both came from so the placement can be judged.
The second simulation, on designing and characterising a superconducting cat qubit, is real work in France and the subject of earlier pieces in this chapter. But no admitted document in this register names superconducting qubit design as a work step or a shortage. A competence list that names a physics topic is not a job description, and a company press release is not evidence of a shortage. We print it as unplaced and keep it where the earlier pieces placed it, against the hardware they documented.
Five jobs CareerOn cannot yet rehearse
The more useful finding for our own product is what the evidence names and we do not cover. The Commission names three shortages we cannot rehearse: quantum software engineering, system integration and quantum cybersecurity. The competence framework names two operational profiles beside the "‘Classical’ engineer with QT add-on": the "QT lab technician (e.g. for operation and maintenance)" and the "QT assembly and test technician." Both technician profiles matter in France for a reason the earlier pieces documented: the machines the State buys have to be installed, kept cold, tested and kept running, and that work is done by technicians as much as by doctors of physics.
Five jobs the evidence names that CareerOn cannot yet rehearse
| Profile | Named as | Publisher, date |
|---|---|---|
| Quantum software engineer | A most critical shortage | European Commission, July 2025 |
| System integration engineer | A most critical shortage | European Commission, July 2025 |
| Quantum cybersecurity specialist | A most critical shortage | European Commission, July 2025 |
| Quantum lab technician | A qualification profile | Quantum Flagship, April 2024 |
| Quantum assembly and test technician | A qualification profile | Quantum Flagship, April 2024 |
Sources 4 European Commission · 5 European Union
Each of these five is recorded as a finding for the simulations catalogue, with the document that names it. We have not stretched a neighbouring simulation to cover any of them. A candidate preparing for a software role in a quantum company should know that our cryogenic simulation is not that preparation, and we would rather say so than let a link imply it.
What this means for your career
If you come from physics, the documents give you a direction rather than a warning. The audit's complaint is not that physicists are too many; it is that training stops at physics. The needs it names, architecture, error correction and the link between classical and quantum computing, sit between a physics degree and a software one. A physicist who can write production code against a real machine, or who understands how a classical control system schedules and corrects a quantum one, is standing in the part of the field every ranking document calls short.
If you come from software, the evidence is the plainest it gets in this chapter. One company's representative said the hard hire was the software engineer, and the Commission ranks quantum software engineering among its most critical shortages. What the documents do not say is that a general software background is enough. The framework's "‘Classical’ engineer with QT add-on" is the profile to read: an engineer from another field who has learned enough quantum to work on the machine. That add-on is the thing to build.
If you come from electronics, you are named more often than any other engineer in the interview study, and you are the one profile CareerOn can help you rehearse now. The simulation Cryogenic electronics engineer: integrated quantum control at 4 K puts you on the control side of the machine, the work the framework describes as improving the control electronics for a qubit realisation. Treat it as one concrete way into that work, read together with the two documents that name it.
If you are drawn to the superconducting hardware itself, the simulation Superconducting qubit engineer: designing and characterising a cat qubit remains in the catalogue, and the earlier pieces of this chapter describe the companies doing that work. In this register no admitted document names it as a shortage, so we print it as unplaced here and make no claim that rehearsing it answers the gap these documents describe.
If you want to be a technician, the framework names your profiles and the French recommendations ask for training aimed at you, and CareerOn has no simulation for you yet. That is our gap, not yours, and it is recorded.
What we could not prove
We found no admitted document that counts open quantum roles in France or in the European Union. We found no dated document that sizes the software shortage against the hardware one; the order we print is an order in words. The interview study is 34 conversations, and one start-up quote is one voice. The OECD postings describe three countries outside the Union, and we have not printed them as European demand. The two documents behind our one placement share a programme, which we disclose rather than resolve. If a dated count of open roles is published, we will add it and say what it changes.
This closes the quantum chapter of the Atlas. Across five pieces the chapter followed the money, the hardware, the buildout, the arguments and now the jobs, and in each case printed what a dated document says beside what it measures. The next chapter will be chosen the same way: from documents first, and from the jobs a candidate can rehearse.