Credentials & evidence

Credentials and evidence

Engagements are led by Ranjeet Rana (Ranj) through SECURE-WIRELESS LTD. The evidence below distinguishes individual qualifications, prior career experience and internal research. His career includes 18+ years in technology programme delivery across financial services, telecoms and travel technology.

Certifications · September 2026

Claude Certified Architect — Foundations

Issued by Anthropic via Credly: 27 September 2026 – 27 September 2027. Covers AI system design, context management, multi-agent orchestration, production reliability and tool/MCP design. An individual qualification; the workshop is independently delivered.

Verify this credential ↗ — we publish links, never score reports or candidate IDs.

Claude Certified Associate — Foundations

Issued by Anthropic via Credly: 10 September 2026 – 10 September 2027. Client-facing Claude knowledge: use-case selection, engagement setup and responsible use. An individual qualification.

Verify this credential ↗

Individual credentials belong to individuals. They are not company partner-tier claims, and we make none. Both public badge pages were checked on 30 September 2026.

Prior career experience

Prior programme and governance experience

The roles below describe Ranjeet's prior career experience — not Secure Wireless AI client engagements or endorsements.

Environment	Delivery role
Travelport	Planning and delivery governance on a global platform customer upgrade
Fidelity	Deadline-driven data-centre exit and migration coordination, via a supplier engagement
Lloyds	PMO, change control and financial governance on a major migration programme, via a supplier engagement
BT	Nationwide hardware and operating-system rollout delivery
Barclays and others	Enterprise technology change, supplier and stakeholder governance

Full employment history and references available on request. Recent independent period: applied-AI research and product development through SECURE-WIRELESS LTD, including the work below.

The evidence lab — stated with its limits

Measured research, honestly caveated

Containment research. Prompt-injection and agent-containment experiments run on our own two-node research cluster. Results depend on the models, test conditions and evaluation method used; internal experiments are not evidence of performance in a customer deployment, and we present them as exactly that.

Judge-integrity programme. Ongoing research into whether AI evaluators can be gamed — and how to prevent it — informs the human-review checkpoints we design for clients.

Multi-engine research operations. We run a multi-engine research pipeline with per-result confidence scoring before anything merges — the same discipline we bring to client evidence packs.

Illustrative sample. Sample assessment report — a fictional scenario with synthetic data, demonstrating the report format and method. It is not a customer case study, penetration test, legal opinion or certification.

No inflated claims. Framework mappings (NIST AI RMF, ISO/IEC 42001) are informational. Sample reports are labelled samples. Nothing on this site claims accredited certification we do not hold.