Evidence packets before narrative.
Claims artifacts →AI Architect · multi-agent systems
IqraLearning
Agent systems that make high-stakes decisions explainable.
Atif Jaffery — specialists, orchestration, and hard gates so automation stays defensible when the answer matters.
All work · 8 projects
Systems built for real decisions.
Flagship first, then the full set.
Flagship · Healthcare Technology
AI Claims Platform V2
Event-driven claims decisions you can audit—agents gather evidence, debate conflicts, and a gate stops unsupported denials.
Open case study →AI & Career Tools
JobHunter
Draft-only career system that refuses to invent résumé lines.
Open case study →Education
ExamReady
Reasoning-first certification practice plus a level-aware voice interview panel.
Open case study →Fintech
Trading Desk | Fintech AI Research
Eleven-agent paper desk where RISK outranks the judge.
Open case study →Infrastructure & AI
AI Capacity Management Platform
Capacity answers grounded in tools and live pricing—not chat guesses.
Open case study →Community Technology
Riwaq | Masjid CRM Platform
Multi-tenant mosque operations—with AI that drafts, humans that approve.
Open case study →PropTech
PropIntel NC | Investment Property Intelligence
Distressed-deal intel with cited AI proposals and rule-based math.
Open case study →Marketing Automation
AI Marketing Operations Platform
Campaign operations where AI proposes and humans approve spend and copy.
Open case study →Trust stack
Guardrails before clever prompts.
The same three controls run through every case study: grounded evidence, hard gates, and a record of who decided.
Denial gate, QA reject, RISK veto.
JobHunter QA →Who decided, on what, how conflict resolved.
Claims ATE →How I build
I design the system, write the specs, and review every change.
Implementation runs on AI coding agents, mostly Claude Code, working from my specs and one set of engineering rules. A separate adversarial agent checks every spec and change for gaps before I sign off.
What I own
- Problem framing and unacceptable failures
- Architecture, agent contracts, and human checkpoints
- The engineering rules every agent follows
How implementation runs
- Specs with acceptance criteria any agent can execute
- Adversarial review of every spec and change
- Thin slices, each tested and reviewed
- Docs, change log, tech-debt list and project tasks updated with every change
What I refuse
- Invented metrics or demo theater
- Agents that deny, spend or submit without a gate
- Generated code nobody reviewed
- Calling something built when it’s only designed
Contact
Hiring an AI Architect?
Tell me about the role or the system you are building. Multi-agent systems with explainable decisions — not demo theater.
Prefer email? a•••••••@gmail.com