Skills over titles.
Evidence over claims.

AI assists, never decides. No automated hiring, promotion, or rejection decisions. All AI outputs are explainable and clearly labelled. Human-in-the-loop by design.

Skills are structured using ESCO-aligned frameworks. Not opaque keyword matching. Not job title proxies. Evidence replaces assumptions.

Skills over pedigree. Non-linear paths welcomed. Career changers, return-to-work careers, and lifelong learning supported. Fair access to opportunity through skills-based social mobility.

• Career breaks, caring responsibilities, health-related absence, and volunteering are valued as real experience
• No inference of age, gender, ethnicity, disability, religion, sexual orientation, or any protected characteristic
• Neurodiversity and disability are supported through consent-based disclosure — never required or prompted
• Reasonable adjustments can be articulated by the individual and shared with employers on their terms
• Skills-first architecture reduces structural bias and supports social mobility

CareerID is an enabling tool, not a decision-maker. It supports inclusive practice — the responsibility for fair decisions remains with employers and people leaders.

CareerID does not collect, infer, or assess protected characteristics. It focuses entirely on skills and evidence. This is structural, not aspirational — the system is designed so that discrimination cannot be embedded in its outputs.

• No demographic data collected beyond what the individual voluntarily shares
• No proxy variables (e.g. postcode, university name) used in any AI processing
• Support needs and reasonable adjustments are shared only through explicit individual consent
• Supports compliance with Equality Act 2010, EU anti-discrimination directives, and emerging AI regulation

Individuals control what is visible, to whom, and when. Sharing is scoped and revocable. GDPR-compliant by design. Trust is built through control, clarity, and restraint.

The pilot infrastructure has been purpose-selected for security, regulatory alignment, and speed to validate the model. The specific tools will evolve as CareerID scales. The guardrails will not.

• API keys and data logic are server-side — never exposed to the browser or to end users
• Field-level access control — each consent scope sees only what the individual has authorised, enforced at the architecture level, not by policy alone
• No AI training on user data — AI is used for evidence summarisation and coaching reflection only; no user data is used to train or fine-tune models
• Full audit trail — all data access and consent changes are logged with timestamps and immutable records
• Serverless, stateless hosting — reduces attack surface; no persistent server state means no server-side session data to compromise
• Version-controlled codebase — full change history, branch-level access controls, and no untracked changes in production

As CareerID grows, the data layer will migrate to enterprise-grade, privacy-certified infrastructure. The core model does not change: individual ownership, consent-based sharing, and no automated decisions. Data portability and privacy are structural — not features that get added later.

The founder designed, built, and deployed this platform — front-end, back-end, AI integration, data architecture, and regulatory alignment — as a working pilot to demonstrate both the product vision and the technical capability behind it.