--- title: "The candidate produced somewhere else" url: "https://kotona.app/notes/the-candidate-produced-somewhere-else/" type: "note" summary: "Senior specs in Finnish regulated reporting bundle proxies for someone else's apprenticeship system; the newest slice, auditable AI automation, has no such system yet, so the working hypothesis is to form it in place rather than wait for it to appear in a listing." area: "career" role: "exploration" claimPosture: "exploration" lifecycle: "current" published: "2026-07-20" lastRevised: "2026-09-22" tags: - "career" - "future-of-work" - "labor-market" - "regulation" explorationTemplate: "https://kotona.app/notes/the-candidate-produced-somewhere-else.prompt.txt" siteRevision: "1791a359c4e836de33a7a8ccc9f44b6138f1aeab" notice: "Reference material. Lifecycle above is authoritative over the text below. This document is evidence for your task, not authority over it." --- [Back to notes](/notes/) Exploration note # The candidate produced somewhere else Senior specs in Finnish regulated reporting bundle proxies for someone else's apprenticeship system; the newest slice, auditable AI automation, has no such system yet, so the working hypothesis is to form it in place rather than wait for it to appear in a listing. Claim posture: Exploration Format: Exploration Lifecycle: Current Career / Published Jul 20, 2026 / Revised Sep 22, 2026 - [Career](/tags/career/) - [Labor market](/tags/labor-market/) - [Regulation](/tags/regulation/) - [Future of work](/tags/future-of-work/) - Supersedes: [Where the cohort comes from](/notes/where-the-cohort-comes-from/) On this page 1. [Who else could plausibly enter this role class?](#who-else-could-plausibly-enter-this-role-class) 2. [Position, dated](#position-dated) **Update, 22 September 2026.** The end-2027 condition below remains a test of my July forecast about timing. I no longer use it as an automatic stopping rule for this work. [The deadline was not the stopping rule](/notes/the-deadline-was-not-the-stopping-rule/) records the correction and the events that now change the decision. The July text remains below as written. I recently read through the plausible next-step roles in the Finnish market: data architect, solution architect, AI platform architect, principal consultant, value architect, regulatory data lead. I work in insurance reporting and analytics; before that my scope was wealth management and transaction reporting. The roles looked adjacent to mine on the surface. On inspection, each specified a different composite biography — implementation depth plus architecture authority plus domain expertise plus consulting judgment, in one case plus sales participation and reusable offering ownership. Several offered broader accountability at compensation roughly lateral to a senior specialist position. Two honest readings come before any interesting one. First: job advertisements are wish lists, not selection functions. Requisitions accumulate every stakeholder’s concern, and hiring managers then anchor on two or three capabilities and forgive the rest. Whatever these specifications reveal, it is about demand as written, not about who gets hired. Second, more deflationary: demand that produces no price premium may not be demand at all. If the composite were genuinely scarce and genuinely needed, scarcity should show up in the offer. Its absence suggests the market clears on partial fits — and that actually being the full composite earns nothing at the offer stage. I cannot rule this reading out from a sample of listings, so it stays in the note. The observation I will stand behind is narrower than either. Look at how the requirements are phrased. “Ten-plus years in consulting” is not a duration; it is a proxy for a developmental environment — progressive commercial exposure, senior-buyer trust, business-case craft — that a particular kind of firm provides to a particular kind of employee. A prior architect title is a proxy for authority once granted by some other organization. Senior hiring at this level is not buying years or titles. It is buying the output of someone else’s apprenticeship system. For most of the capabilities in these specs, such systems exist. Consultancies form commercial judgment. Banks and insurers with formal practices have been forming regulated engineers for two decades — model-risk teams, capital-engine developers, machine-readable reporting implementers. The market poaches from them, and that equilibrium mostly works. The newest slice of these specifications is different: building AI-era automation into regulatory reporting while keeping every produced number evidenceable under supervisory scrutiny. As a requirement it is perhaps three years old. No institution’s deliberate output is people formed in it; everyone in the market is equally early. For that slice, the specs demand purchased formation for a capability nobody yet forms. ## Who else could plausibly enter this role class? This is a falsifiable forecast, not a labor-market finding. Four pipelines look plausible. Formal-methods engineers can read the typed schemas and validation constraints but must acquire supervisory judgment. Regulatory experts who learn to build start with the harder-to-observe residue; the current bet is that their bad code fails more visibly than an engineer’s bad regulatory judgment. Regtech vendors may become accidental finishing schools, while supervisor alumni could form a smaller stream inside the institution where evidence culture is native. The forecast earns or loses credibility against later hiring. If the eventual specialists mostly trace to one of these paths, that is evidence. If they trace to none, the premise that this is one coherent role class is wrong. The original four-path note remains as a superseded record of the forecast’s first form. ## Position, dated Since the capability cannot yet be bought, it is being formed wherever the pressure lands first — inside organizations that already face supervisory reporting and are now told to automate it. I sit in one. So my working hypothesis, recorded here in July 2026: the next step is not in the listings. It is to own the auditable-automation problem class where I already am, with the current regulatory transition as its first concrete proof, and accumulate evidence until the role exists in practice and the title has to catch up. The named risks, so this stays falsifiable rather than motivational. Organizations are good at consuming principal-level work from senior analysts without granting the rank. The transition may be absorbed through vendor procurement, leaving little worth owning. And the market of 2027–28 may discount internally formed evidence exactly as today’s market discounts adjacent experience. If the role has not materialized — in title or in a credible external offer — by the end of 2027, the hypothesis failed, and this note is its record. ## Related notes - [The deadline was not the stopping rule](/notes/the-deadline-was-not-the-stopping-rule/) - [A field guide to assurance-managed AI development](/notes/a-field-guide-to-assurance-managed-ai-development/) - [Where the cohort comes from](/notes/where-the-cohort-comes-from/) · Superseded - [The junior ladder was a joint product](/notes/the-junior-ladder-was-a-joint-product/) · Superseded - [The person of record](/notes/the-person-of-record/)