Product · AI systems · Regulated finance · Riyadh
I’ve built the products, ventures and digital banks that regulated finance runs on.
After 20 years building, I’m focused on AI decision systems that hold up inside institutions — where policy, data, model output and human judgment stay separated, logged and auditable. The goal is institutional intelligence: faster, more consistent decisions that can be inspected, governed and trusted when the stakes are high. That is the work here.
/ Selected Work
Personal builds on public and synthetic data — designed as regulated-finance systems.
Sentinel
A geopolitical risk engine for committee-ready exposure briefs. Given a Strait-of-Hormuz oil disruption, Sentinel identifies exposed positions, ranks materiality and traces each figure back to source disclosures — so the committee acts on evidence, not a model score.
Mizan
A forward-tenor affordability engine for Saudi BNPL. When bureau data lags real-time borrowing, Mizan reconstructs hidden obligations from bank-account signals, tests whether the customer can carry the full repayment term and returns a right-sized approval or counter-offer — before a loan that clears policy becomes unaffordable months later.
Fatoora
A receivables engine for thin-file SMEs. Fatoora underwrites the buyer behind the invoice by matching paid invoices against the cash that actually landed, scoring buyer settlement behaviour and flagging fabricated, duplicate or over-concentrated receivables at intake.
Exploring (3)ShowHide+−
Trade-finance AML analyst
Reads a documentary credit for trade-based money-laundering patterns — over-invoicing, circular shipments, fronted counterparties — and attaches the relevant red flag or typology behind each alert, so escalation is evidence-led rather than score-led.
Shariah-audit copilot
Tests murabaha, ijara and tawarruq documentation against approved product structures, drafts the evidence basis for Shariah review and flags where execution appears form-compliant but economically misaligned with the intended structure.
Deposit-stability model
Models deposit behaviour under stress before it appears in liquidity ratios: which non-maturity balances are operational, rate-sensitive or flight-prone; which accounts run off; and how quickly liquidity leaves under scenario pressure.
/ Contact
If you’re building where AI meets regulated finance in the GCC, I’d value the conversation.