Edge-Data Productisation Architect — Offshore Wind Operations

The mandate

The provider collects vibration, temperature, controller and maintenance signals from several turbine generations, but each analytics team creates its own extraction, resampling and quality rules. Promising condition-monitoring models therefore require costly site-specific integration and cannot explain when missing edge data invalidates a recommendation. Product leadership wants reusable maintenance data products without forcing operators to surrender raw control data or adopt one cloud.

The named deliverable is an Edge-Data Productisation Kit for gearbox and converter maintenance, comprising semantic contracts, edge transformations, quality and confidence rules, synchronisation, cyber boundaries, product service levels, operator rights and commercial cost. Reference implementations must operate across two turbine generations and disconnected periods, with interfaces portable across an operator-managed and provider-managed deployment.

Milestone one at week three delivers decision journeys, source variation and product-boundary options. By week seven, milestone two supplies contracts and a working edge reference for one subsystem. Week twelve concludes field trials across both generations and two connectivity conditions. In week seventeen, milestone four provides accepted product kits, economics, implementation backlog and a trained internal product council.

Acceptance requires field engineers to use the products for six unseen maintenance cases, trace each recommendation to source confidence and identify when action must be deferred. Cyber must approve edge and transfer boundaries; two operator architects must demonstrate interface portability without provider-only tooling; and Finance must reperform service cost. A successful analytics demo without field decision adoption does not satisfy acceptance.

The client will provide representative controller histories, sensor mappings, maintenance records, edge hardware, connectivity profiles, cyber policies, customer contracts and product cost. Operators will nominate field engineers and architecture contacts under data agreements. Client teams will implement privileged controller access, while the product chief will resolve disputes over raw data, derived rights and support commitments within three working days.

Why this is external work

Analytics teams favour feature access, edge engineers favour hardware constraints and commercial teams promise customer flexibility; none can independently set a durable product boundary. Operators also distrust architectures that appear to create provider lock-in. An external specialist can connect maintenance consequence, industrial data engineering and portability while remaining neutral about the eventual cloud or edge vendor.

What you will own

• Map gearbox and converter maintenance decisions to required signals, temporal context, calibration, quality, confidence and evidence-retention needs.
• Define edge transformations and semantic contracts that remain consistent across turbine generations without concealing source-specific limitations.
• Design synchronisation, correction and lineage behaviour for disconnected operation, late events, controller replacement and changed sensor mapping.
• Establish cyber and rights boundaries among raw control data, derived condition features, product outputs, operator history and cross-customer learning.
• Trial the products through unseen field cases where missingness, conflicting evidence or low confidence should alter or defer maintenance action.
• Build service economics across edge compute, data transfer, integration, support, model dependency and turbine-generation variation.
• Transfer product qualification, contract change, field feedback and portability tests to an internal council spanning product, engineering and operations.

Candidate qualifications

• Productised edge or industrial data for wind, power, heavy equipment, aviation or another safety-conscious asset-maintenance environment.
• Designed semantic and quality contracts across equipment generations whose controllers, sensors and connectivity could not be standardised retroactively.
• Linked data confidence to real maintenance decisions and can evidence when a system correctly withheld or qualified its recommendation.
• Negotiated raw, derived and learned-data rights between equipment providers and asset operators without treating cloud centralisation as inevitable.
• Validated disconnected synchronisation, edge transformation and portable interfaces through field use rather than laboratory simulation alone.
• Built full data-product economics including integration and support burden, then transferred qualification to permanent product teams.

Non-negotiables

• The named architect must attend Helsinki design sessions and North Sea field trials, including travel governed by offshore safety requirements.
• No reseller, referral or investment interest may exist in edge hardware, cloud, turbine analytics or industrial platform suppliers considered.
• The engagement cannot modify safety-critical turbine controls or treat data-product acceptance as approval of a maintenance decision system.
• Operator data rights, regional transfer constraints and field-safety authority must remain explicit in every reference implementation.

1. 49 words maximum. Describe an edge-data product that correctly withheld an asset-maintenance recommendation because its input confidence was inadequate.
2. 49 words maximum. How would you preserve one semantic contract across turbine generations with materially different sensor mappings?
3. 49 words maximum. Which field test would demonstrate portability between operator-managed and provider-managed edge deployments?