Engineering Knowledge-Recovery Leader

The mandate

A new enterprise-search layer indexed millions of documents but made engineering judgement harder to recover. Results privilege recent presentations over approved test evidence, component names vary by laboratory and the reasons behind design exceptions live in expert memory. Simultaneously, several specialists are retiring. The transformation sponsor resigned after teams duplicated a thermal-management decision already resolved elsewhere, creating a fourteen-month leadership gap with direct product-cycle consequences.

The interim leader will reorganise knowledge around engineering questions, governed entities and decision provenance rather than another content migration. The mandate covers component identity, requirement and test relationships, design rationale, supersession, confidentiality, expert contribution and retrieval evaluation. Search, graph and generative interfaces may support the solution, but every answer used for design must reveal source authority, applicability, revision and known disagreement.

Permanent knowledge-product and engineering-stewardship leaders will be appointed by month nine. They must jointly own the final laboratory rollout, resolve a disputed supersession case and present reuse evidence at the global engineering council before handover. Transfer includes the domain model, evaluation corpus, stewardship load, security boundaries, vendor dependencies, adoption economics and unresolved ownership choices, plus named specialist succession plans.

Within the approved envelope, the seat can stop ingestion of unclassified material, prioritise domains, define retrieval quality gates, appoint engineering stewards, redirect platform capacity and require decision records for designated reviews. It cannot approve product designs, alter retention policy, declassify research, select components, waive export controls, accept model-generated evidence or commit laboratories to headcount changes outside delegated authority.

The remit excludes acting as design authority, corporate archivist, information-security officer or permanent search product owner. Success means engineers find applicable rationale and evidence faster, duplicated analysis falls, retiring expertise is transferred into governed decisions and successors can operate domain stewardship without a central rescue team. Search usage alone is not an outcome if retrieved material remains contextless or untrusted.

Why this seat is open

The previous sponsor treated indexing breadth as transformation progress and could not resolve who owned engineering meaning once content crossed laboratory boundaries. A costly duplicated decision made that failure visible as senior experts approached retirement. The company needs temporary executive authority to connect platform capability with disciplined engineering stewardship, protect scarce knowledge now and shape permanent roles around observed product decisions.

What you will own

• Map priority engineering questions to components, requirements, tests, decisions, exceptions, experts and authoritative source systems.
• Establish identity, version, applicability, supersession and disagreement rules across laboratories using different technical vocabularies.
• Build a retrieval evaluation corpus from real design questions, misleading near-matches, obsolete evidence and controlled-access cases.
• Define decision-rationale capture that fits engineering reviews and records alternatives, assumptions, evidence, authority and expiry triggers.
• Govern security, export-control and need-to-know boundaries across indexing, retrieval, summaries, expert discovery and usage telemetry.
• Direct laboratory adoption around live reuse decisions and measure search success through avoided duplication and decision quality.
• Qualify permanent leaders through rollout and contested stewardship choices and transfer the platform economics and operating calendar.

Candidate qualifications

• Has transformed enterprise search or knowledge systems inside complex hardware, electronics or engineering-led product development.
• Can model components, requirements, tests, decisions and supersession so retrieval preserves technical applicability and provenance.
• Has built evaluation sets that expose plausible but obsolete or contextually wrong results rather than rewarding keyword relevance.
• Understands laboratory confidentiality, export controls, research retention and why generative summaries cannot become design evidence.
• Has embedded stewardship into engineering work without creating a separate documentation bureaucracy that specialists bypass.
• Demonstrates succession and measurable component reuse after expert departure, not only platform launch or content-volume growth.

Non-negotiables

• Will work onsite in Seoul and complete monthly laboratory residencies and quarterly global engineering councils.
• Must disclose relationships with electronics manufacturers, search vendors, knowledge-platform suppliers and engineering consultancies.
• Brings knowledge recovery tied to real engineering decisions; generic intranet, collaboration or records programmes are insufficient.
• Will not permit generated text, popularity or recency to substitute for attributable and applicable design evidence.

1. 49 words maximum. Which retrieval result looked relevant but was unsafe because its engineering context had expired?
2. 49 words maximum. How would you capture design rationale from a retiring specialist without creating a memoir project?
3. 49 words maximum. What decision must the permanent stewardship leaders resolve together before handover?