The infrastructure job market values recent certifications, fashionable technologies, the ability to talk Kubernetes or Terraform in a 45-minute interview. It values poorly — often not at all — 20 years of institutional memory, the diagnostic capacity under pressure built on thousands of real incidents, and the trust network with other seniors capable of solving problems that runbooks do not cover. These competences do not fit in a technology list on a CV. They do not pass ATS filters. They have no associated certification.

The more tacit knowledge the senior accumulates, the less visible they are in the market. At 45, their certification list is often older than that of a recently trained junior. Their salary is higher — which makes them a cost-reduction target during reorganisations. Their knowledge is deep but invisible — which makes their contribution difficult to justify in a dashboard. The sector thus creates a mechanism of progressive expulsion of its most experienced profiles, precisely when their knowledge reaches its maximum maturity.

The more irreplaceable the senior is on legacy systems, the less the organisation can afford to release them to build skills on new systems. The longer they stay on legacy, the more indispensable they become on legacy and the more obsolete on the modern stack. The organisation thus creates a human technical debt — an expert on critical systems who cannot evolve because their evolution would endanger the continuity of the systems they maintain. This debt accumulates silently until the day the senior leaves — and the organisation simultaneously discovers what it had and what it no longer has.

The standard organisational response to a senior's departure is the "handover" — a few weeks of intensive documentation before leaving. This response is structurally insufficient. Tacit knowledge is not the knowledge the senior knows they have. It is the knowledge they mobilise without knowing it — automatic diagnostic patterns, intuitions about abnormal system behaviour, the memory of the contexts in which a decision was made. This knowledge cannot be documented because it is not conscious. It transfers through shared practice, through apprenticeship, through years of working side by side.

No organisation calculates the true cost of an experienced senior's departure. The cost of replacement is calculated — recruitment, the successor's salary, the training period. What is not calculated is the cost of lengthened incidents, of poorly-made decisions for lack of historical context, of clients lost because a problem was not resolved fast enough, of months spent bringing the successor up to speed on systems they do not yet understand. This invisible cost is systematically higher than the visible cost of recruitment — and nobody sees it coming.

Successive reorganisations dissolve the stable teams where the senior built their trust network and informal legitimacy. In a new team with a new manager who does not know their history, the senior starts from scratch in building their credibility — while their years of experience remain invisible to a manager evaluating them on the same criteria as their junior colleagues.

Annual reviews and skills assessments are built around fashionable technologies. A senior who perfectly masters critical production systems but does not know the latest container orchestration tool will be negatively assessed against a skills framework designed for juniors in training. The evaluation measures what the senior does not yet know — not what they have known for twenty years and that no one else knows.

The senior is the team's most expensive employee. Under payroll pressure, they are the obvious target for budget optimisation. The argument presented is always economic — "reorganisation", "strategic evolution", "rightsizing". The reality is that the senior costs a lot and their cost is visible, while their value is not. The organisation systematically arbitrates in favour of the visible against the invisible.

Twenty years of on-call, twenty years of 3am incidents, twenty years of responsibility for critical systems without that responsibility being recognised at its true value — the senior who does not leave through the economic door often leaves through the exhaustion door. This burnout is not a sudden event. It is a silent accumulation that spreads over years, in a culture that values resilience under adversity without ever questioning the structure that produces that adversity. This mechanism will be examined in depth in Angle 4 — "The Pager and the Body".

Tacit knowledge transfer has only one effective vector: shared work on real problems. A senior and a junior working side by side on the same incidents for 12 to 18 months transfer more knowledge than intensive documentation over 3 months before departure. This pairing must be structural — not occasional, not reserved for crisis moments. It must be integrated into the organisation of work as a permanent practice, not an emergency response.

The senior's cognitive desertification is avoidable if the organisation guarantees exploration time — protected, regular, non-interruptible by production incidents. 20% of working time dedicated to technological exploration is a documented practice in some organisations. It is uncommon in infrastructure. It should be standard — because a senior who stays current is a senior who stays relevant, and a relevant senior is a senior the organisation has no reason to expel.

The senior must have a formal role that recognises their tacit knowledge and gives them an organisational legitimacy independent of their mastery of recent technologies. Titles like "reference architect", "domain technical expert", or "critical infrastructure lead" are not honours — they are retention and valuation mechanisms that make visible what was invisible. They also protect the senior from reorganisations that reduce experience to a cost and ignore its value.

Every experienced senior should have, at any time, a documented transmission plan — not an emergency handover, but a continuous process of documentation and transmission that spans several years. This plan is not a threat ("you are going to leave") — it is an organisational insurance policy ("when you leave, whether in 1 year or 10, the team will not be blind"). This practice is standard in other critical sectors. It is almost absent in digital infrastructure.

Every organisation should calculate, after each experienced senior departure, the total cost of that departure — successor recruitment, lengthened incidents in the first 12 months, successor training and support time, external consultants mobilised. This calculation, made visible to management and HR teams, radically changes the cost/value ratio of the senior. A senior at €80,000 annually whose departure costs the organisation €300,000 is no longer expensive — they are underpaid.

The skills frameworks used for annual reviews must integrate tacit knowledge criteria: institutional memory, diagnostic capability on legacy systems, transmission to juniors, off-runbook incident resolution. These criteria are no less objective than a certification list — they simply require a different evaluation mode (observation in real situations, peer feedback, incident history). Their absence from current frameworks is a choice, not an inevitability.

A formal status recognised in collective agreements, which protects the senior against reorganisations — not for union reasons, but for service continuity reasons. This status is assigned on objective criteria (seniority, scope of responsibility, critical systems mastered), is reviewable, and comes with an obligation of documented transmission. It makes tacit knowledge visible in the organisational chart.