Enterprise AI.
Built for Trust.
Practical, governed AI for regulated enterprises – with security, compliance and measurable value designed in from the start.
All projects use synthetic or public data. No client or confidential information is used.
Governed
Enterprise IT
Governed
Enterprise IT
Governed
Enterprise IT
Governed
Enterprise IT
Governed
Enterprise IT
PORTFOLIO
ENTERPRISE AI PORTFOLIO
Practical AI solutions demonstrating how enterprise AI can be designed, secured and governed across regulated industries.
- All
- AI GOVERNANCE NEXUS
- CALIBRE
- SECURE-RAG
- SURVEILLANCE
- AGENTIC RISK ANALYST
FLAGSHIP • ENTERPRISE AI GOVERNANCE
AI Governance Nexus
One governance layer for the AI lifecycle.
AI Governance Nexus brings risk assessment, governance classification and evidence-grounded knowledge into a single lifecycle-aware platform for governing AI initiatives from Proposed to Retired.
What it demonstrates
♦ AI lifecycle orchestration – connects Agentic Risk Analyst, CALIBRE and Secure-RAG.
♦ Deterministic decision engine – converts structured findings into auditable executive decisions.
♦ Persistent AI governance – maintains initiative history, assessments, decisions, controls and review dates.
Key capabilities:Initiative Register, Decision Engine, Trust Score, Evidence Pack, Executive Decision
Architecture: FastAPI, Python, SQLite, HTMX, LLM, RAG
LIVE APPLICATION →
AI GOVERNANCE • RISK QUANTIFICATION
CALIBRE
Measure how much governance an AI system actually needs.
CALIBRE is a transparent AI governance framework that converts five characteristics of an AI system into a Governance Intensity Index, governance band and corresponding enterprise controls.
What it demonstrates
♦ Explainable risk scoring – every score can be traced to its underlying factors.
♦ Proportionate governance – avoids over-governing low-risk AI while escalating consequential systems.
♦ Control mapping – connects governance requirements with NIST AI RMF, ISO/IEC 42001 and EU AI Act references.
Five assessment dimensions: Autonomy, Consequence, Irreversibility, Data Sensitivity, Population Scale
Outputs: 0–100 Governance Score, Risk Band, Control Set, Approval Route, Evidence Pack
LIVE APPLICATION →
GENERATIVE AI • RAG • AI SECURITY
Secure-RAG
Enterprise AI answers grounded in evidence – not guesswork.
Secure-RAG provides evidence-grounded answers to regulatory, policy and enterprise knowledge questions, with every answer constrained by retrieved source material and an explicit “insufficient evidence” path.
What it demonstrates
♦Grounded generation – answers are generated from retrieved authoritative passages.
♦ Hallucination controls – citation requirements, confidence classification and refusal when evidence is insufficient.
♦ Traceable knowledge – every answer can be traced to the underlying document and source passage.
AI pipeline: Embedding → Retrieval → Evidence Scoring → Grounded Generation → Citation
Technology: RAG, BGE-M3, Hugging Face, LLM, FastAPI
LIVE APPLICATION →
AGENTIC AI • RISK INTELLIGENCE
Agentic Risk Analyst
From an AI idea to a structured risk decision.
Agentic Risk Analyst uses specialist AI agents to analyse proposed or existing AI initiatives across business risk, regulatory obligations, cybersecurity and controls, producing a structured assessment for executive review.
What it demonstrates
♦ Multi-agent reasoning – specialist Risk, Compliance and Security agents work in parallel.
♦ Structured AI outputs – typed JSON contracts make findings machine-readable and reusable.
♦ Risk-to-control reasoning – connects inherent risk to controls, residual risk and remediation actions.
Assessment pipeline: Scenario Intelligence → Risk + Compliance + Security → Controls → Residual Risk → Executive Decision
Technology: Multi-Agent AI LangGraph LLMs MITRE ATLAS NIST AI RMF Structured JSON
Problem: Maritime operators need to identify and assess potentially unauthorised vessels in large surveillance areas.
Solution: AI-powered computer vision for vessel detection, classification and risk assessment.
Releasing Soon
Problem: Aged-care environments need timely detection of incidents while respecting privacy and maintaining human oversight.
Solution: Edge AI and computer vision to detect potential incidents and provide timely alerts to caregivers.
Releasing Soon
Security by Design
PII protection, access control and audit built into every project, not added later.
Governance wrapper
Every model incorporates explainability, human oversight and evidence on demand.
Synthetic data only
Public and synthetic datasets only. No client or confidential data.
Cloud-agnostic
Open-source foundations, deployed on Azure, AWS or other platforms where they add value.
Need governed AI in your enterprise?
From financial crime and cybersecurity to legacy modernisation and AI governance, I help regulated organisations adopt AI securely and at scale.