By the end of the first week, Shine had attended at least one session for each of her courses in the Drug Regulatory Affairs and Policy programme. So far, the courses seemed decent. Nothing in those first sessions justified dismissing the teaching itself, and there were signs that the programme could offer useful regulatory knowledge.
But this did not ease her distress. The question troubling her was no longer simply whether the courses would be good. It was whether completing them would provide enough practical experience, credible evidence and access for her to obtain the kind of Regulatory Affairs job the programme was meant to prepare her for. A course could be worthwhile without becoming the missing bridge between a student and a profession.
There was one reason not to treat the present market as permanent. The Hong Kong Government plans to establish the Hong Kong Centre for Medical Products Regulation, or CMPR, by the end of 2026. If the new institution creates regulatory work and strengthens the surrounding ecosystem quickly enough, Shine may encounter opportunities even before she graduates. But an institution’s establishment is not a promise to an individual student. It does not guarantee suitable junior positions, eligibility for them or a timetable that matches her graduation.
The semester had only just begun. Shine was still much closer to “knowing almost nothing” than to feeling like a finished Regulatory Affairs professional, yet graduate recruitment was already under way. The calendar created its own instruction: ready or not, Shine had to enter the job market now.
Shine was sending her CV through every plausible channel she could find, then returning from each round of searching with the same result: employers wanted senior Regulatory Affairs professionals.
Eli Lilly was recruiting a senior RA again. IQVIA offered another version of the same answer. Even the genuinely junior, low-paid position Shine eventually found appeared to require a credential designed for people who already had professional experience.
The individual vacancies were not the whole story. The question was where the next generation of experienced professionals was supposed to come from if employers repeatedly removed or devalued the stage at which people were still learning.
What Shine brought in
Shine was searching for Regulatory Affairs work again. A multinational pharmaceutical company wanted a senior candidate. Then a search of vacancies at a global clinical-research company revealed another senior role.
Neither advertisement was surprising by itself. Regulatory work can carry substantial responsibility: an employer may need someone who already understands local requirements, can maintain registrations, handle authority interactions, coordinate submissions and make decisions without close supervision.
What disturbed her was the repetition.
Large companies appeared to want people who had already become internationally capable Regulatory Affairs professionals. Small companies often did not have enough continuous international regulatory work to employ such a person at all. They might stay within one market, combine Regulatory Affairs with Quality Assurance or general operations, rely on a distributor, or purchase regulatory support only when a project required it.
Shine was being urged at home to distribute her CV as widely as possible, but application volume assumes that suitable openings exist. If the visible market contains mainly senior positions, sending the same junior CV into it more frequently does not create a junior pathway.
“Everyone wants the mature, stable thing,” she said, “but no one wants to be the bridge.”
The first interpretation
When the feeling was close, the market seemed to be delivering a personal verdict: there is no place for someone at her stage.
That interpretation was understandable, but too sweeping. A vacancy describes a company’s immediate hiring decision. It does not measure the total potential of every person excluded by its requirements.
The more useful question was structural:
If employers repeatedly recruit people who already possess five or ten years of experience, where are those people now supposed to acquire their first year?
This is not merely the familiar complaint that entry-level jobs ask for too much experience. It is a question about whether the career ladder still contains its lower rungs.
A junior role asking for proof of maturity
Later, Shine found a small company advertising what appeared to be a genuinely junior, relatively low-paid Regulatory Affairs position. The vacancy nevertheless required the Regulatory Affairs Certification, or RAC.
If the advertisement meant the professional certification administered by the Regulatory Affairs Professionals Society, the conditions did not fit together. RAPS currently requires a bachelor’s degree plus three years of regulatory-related experience, a master’s degree plus two years, or a doctorate plus one year before a candidate may sit the RAC examination.
A recent graduate can, of course, already possess professional experience. But a position that presents itself as an entry point, offers junior-level pay and requires a credential built around applied experience asks the candidate to arrive beyond the stage the job claims to serve.
The inconsistency does not prove that the company intended to deceive applicants. “RAC” might have been copied from another vacancy, confused with a course certificate or included as an ideal preference without careful wording. What it does show is that the job description was not reliable enough to be treated as a precise measurement of Shine’s qualifications.
The large employers wanted finished professionals. The small employer appeared to want evidence of maturity at a junior price.
Is this the general problem of the AI age?
Shine’s broader hypothesis was that artificial intelligence makes this temptation stronger. A company may imagine that it can eliminate routine junior work, retain a smaller number of experienced people, equip them with automation and offer enough money to attract the senior talent it still needs.
There is now evidence that AI is affecting some entry-level hiring. In a 2026 Gartner survey, 22% of chief human-resources officers said that at least one business leader in their organisation had stopped hiring for certain entry-level roles because of AI automation.PwC’s 2026 global analysis described a two-track entry-level market and found that employers increasingly expect junior workers in AI-exposed roles to bring skills traditionally associated with more senior work, including judgement and leadership.
Research using US administrative data has also found a marked decline in early-career hiring in highly AI-exposed industries after large language models became widely available.
But the causal story is not settled. A UK government and LinkedIn analysis found broad weakness in entry-level hiring, while also noting that junior and senior hiring had declined in parallel. The occupations with the sharpest declines included some in which AI capabilities are visible, but the report explicitly concluded that more research was needed before assigning causation.Economic conditions, post-pandemic adjustment, interest rates, remote work, industry restructuring and changes in the supply of graduates may all affect the same numbers.
Nor does the available evidence establish that AI produced the senior-heavy Regulatory Affairs vacancies Shine encountered in Hong Kong. Local market structure offers older and more direct explanations: small affiliate teams, limited headcount, outsourced work, a relatively small pharmaceutical market and the concentration of regional functions elsewhere.
The AI claim should therefore remain a question rather than a conclusion:
Is AI creating the shortage of junior pathways, or is it giving employers a new way to intensify a preference for ready-made experience that already existed?
The work that disappears is also training
Entry-level tasks are easy to describe as routine. A junior employee checks a document, enters information, assembles a first draft, searches a database, updates a tracker or prepares a submission component for review. If a machine can perform part of that work faster, removing the junior position can look efficient.
But each task produces two outputs. The visible output is the completed document or updated system. The less visible output is a person learning what matters, what can go wrong and why a reviewer changes the work.
Automation may preserve the first output while reducing the second.
This does not mean that organisations must preserve inefficient tasks unchanged as educational exercises. It means that if the old apprenticeship mechanism is removed, another one must be designed deliberately. Otherwise, today’s senior employee becomes more productive while the supply of tomorrow’s senior employees quietly contracts.
The contradiction appears only later. Companies compete for the remaining experienced people, salaries rise for selected roles, vacancies stay open and employers describe a talent shortage. Yet the shortage is partly maintained by the same reluctance to employ and supervise beginners.
What the evidence permits
Shine’s search supports a modest conclusion: the Hong Kong vacancies visible to her currently skew towards senior positions rather than the entry-level work she needs. It does not prove that Hong Kong has no junior Regulatory Affairs work, that every company has stopped training or that AI caused the pattern.
The planned establishment of CMPR means that this market snapshot may change. It gives Shine a concrete reason to remain open to opportunities emerging before graduation, but the government timetable states only when the institution is expected to open. It does not establish what junior roles will follow, how quickly the wider market will respond or whether Shine will be eligible for a particular position.
The broader evidence permits another modest conclusion: pressure on early-career work is real in several labour markets, and some employers explicitly connect reduced entry-level hiring with AI. It does not permit the claim that all junior work is disappearing or that one technology explains every weak career ladder.
What remains persuasive is the underlying arithmetic. A profession cannot recruit only mature professionals indefinitely. Experience is not a natural resource waiting to be extracted. It has to be produced inside institutions.
When the senior-only standard moves inward
Near the end of our conversation, Shine recognised the same preference operating inside herself.
She had begun comparing her unfinished writing with the polished language an AI system could produce after receiving her ideas. Her notes looked informal, incomplete and visibly uncertain. The generated version looked composed and ready to present. Because another polished version could be produced quickly, losing one of her own drafts no longer seemed important.
Then she connected this judgement to the vacancies.
The senior-only employer treats the learner as an inferior version of the professional it actually wants. Shine had started treating her own learning in the same way. She found it difficult to tolerate herself while she was still becoming competent—and, under competitive pressure, sometimes found it difficult to tolerate the visible immaturity of other students as well.
This is how a labour-market preference can become a personal standard. Finished performance appears to be the only form worthy of patience. The process that produces it becomes embarrassing, inefficient or disposable.
A junior professional is not a defective senior, just as a draft is not a defective final document. A Phase I study is not a failed Phase III study. Each exists to do work that the later stage cannot perform for itself. An exploratory study identifies what should be tested next; a draft preserves uncertainty and records the movement by which an idea was discovered; a learner performs tasks through which judgement is gradually formed.
Polish can also conceal defects. A fluent AI-generated answer may close gaps that should have remained visible, smooth over weak reasoning or make an interpretation appear more settled than the evidence permits. Shine’s rough formulation may be less elegant, yet it may contain the original connection that made the polished version possible.
The personal problem is therefore not simply low confidence. It is an internalised development policy: accept only mature outputs, while providing no protected space in which immature work can survive long enough to mature.
Changing that policy does not require pretending that every draft or beginner’s effort is excellent. It requires judging developmental work according to its stage and purpose. The source note and the refined article can both be preserved: one as the record of discovery, the other as the communication product. Students can be corrected without being treated as failed professionals. Shine can expect herself to learn without regarding the fact that she is learning as evidence against her worth.
What belongs to Shine
Shine cannot manufacture a junior vacancy by distributing more copies of her CV. She also cannot carry responsibility for a market that wants experience without offering enough supervised work through which to gain it.
What she can do is distinguish several questions that anxiety tends to collapse into one:
- Is this a genuine entry route, or merely a senior role with flexible wording?
- Does an adjacent position provide real regulatory exposure, or only borrow the word “regulatory”?
- Will a small company develop international experience, or combine unrelated functions under one title?
- Could a graduate programme, internship, Regulatory Operations team, consultancy or CRO function as an apprenticeship pathway?
- Is Hong Kong the appropriate place to study and build relationships even if the first permanent Regulatory Affairs role begins in another market?
These questions do not make the missing rung reappear. They prevent its absence from being misread as evidence that she personally failed to climb it.