Closed-Loop Bioelectronic Control Evidence Architect — Chronic Care

The mandate

The development programme is moving from clinician-set stimulation schedules toward controllers that adapt within prescribed limits using physiological signals. Device, sensor and software records exist, yet teams cannot consistently reconstruct whether a changed dose arose from patient state, artefact handling, control logic, clinician instruction or a manual intervention. That ambiguity weakens both safety reasoning and meaningful clinical review.

The engagement deliverable is a Closed-Loop Bioelectronic Control Evidence Architecture covering prescribed intent, device and electrode configuration, physiological observation, signal-quality treatment, inferred state, controller version, constraint, stimulation action, patient response, clinician review, override and correction. It will define two reference therapy patterns without recommending treatment or validating a medical endpoint.

Milestone one at week three produces clinical-control journeys, hazard boundaries and evidence fractures. Week seven concludes milestone two with the closed-loop state model and decision-rights map. At week twelve, milestone three delivers reference traces and controller-degradation exercises. The accepted architecture, supplier schedules, migration priorities and client-run qualification pack close milestone four at week seventeen.

Acceptance requires clinical and engineering owners to reconstruct twelve unseen stimulation episodes to the prescription, usable signal, controller state, enforced limits and human intervention; injected artefact and interrupted sensing must cause bounded behaviour; and independent Quality review must reperform traceability. The medical officer signs after client teams run a constraint-change and delayed-review scenario without consultants.

The client will provide de-identified device and session histories, protocol and prescription structures, controller and firmware manifests, signal-quality rules, intervention records, hazard analyses, supplier terms and controlled test environments. Client engineers build references and clinicians interpret patient evidence. The assignment excludes treatment decisions, clinical endpoint validation, production code, trial operation, regulatory submission authorship and device certification.

Why this is external work

Control engineers understand adaptive logic and clinicians understand therapeutic intent, but each discipline reads a different account of the same stimulation episode. Existing suppliers also own pieces of the evidence chain. External architecture can create a reviewable boundary among sensing, computation and care without designing treatment or issuing a safety or regulatory opinion.

What you will own

• Map prescription, device, electrode, observation, signal-quality treatment, inferred state, controller constraint, stimulation, response and intervention.
• Define effective configuration for firmware, controller, patient-specific limits, sensing channels and clinician instructions across sessions.
• Design artefact and missing-signal behaviour that stays within explicit constraints and makes degradation visible to authorised users.
• Exercise motion artefact, electrode change, sensor dropout, stale prescription, controller rollback and delayed clinical review.
• Specify explanations showing which observed evidence and limits produced an adaptive action without implying medical causality.
• Assess component boundaries through determinism, latency, fail-safe state, provenance, cyber trust, portability and supplier exit.
• Transfer trace review, constraint-change qualification and degraded-loop exercises to permanent clinical and engineering owners.

Candidate qualifications

• Led closed-loop implantable, neuromodulation or bioelectronic control architecture across clinically governed device programmes.
• Connected physiological sensing, artefact processing, patient-specific limits, control logic, stimulation and clinician intervention in auditable traces.
• Designed safe degradation for missing or unreliable signals without silently converting adaptive therapy into uncontrolled fixed behaviour.
• Worked with clinicians, Quality and Regulatory while keeping control-system evidence separate from treatment and endpoint judgement.
• Reconstructed consequential device episodes after configuration change, manual override, delayed review or contradictory physiological signals.
• Delivered supplier-neutral architecture that internal teams exercised against novel artefacts and constraint changes after closure.

Non-negotiables

• The named architect must lead Tel Aviv clinical-engineering workshops and both degraded-loop acceptance exercises.
• No financial relationship may exist with implant, sensor, control-software or clinical-platform suppliers evaluated.
• Clinicians retain prescription, treatment, patient-monitoring and clinical-interpretation decisions throughout the work.
• Production software, trial conduct, regulatory authorship, endpoint validation and device certification are expressly excluded.

1. 49 words maximum. Describe an adaptive stimulation episode whose cause could not be inferred from controller output alone.
2. 49 words maximum. How did you make signal degradation visible while keeping a device inside prescribed limits?
3. 49 words maximum. Which client artefacts are required before testing a stale-prescription scenario?