Portfolio · 2026 · Amman, Jordan
SIEF ALSHEYAB
|
I build production-grade web platforms, AI automation systems, and fintech risk workflows using React, Node.js, Ruby on Rails, Python, FastAPI, Docker, and cloud deployment pipelines.
ReactNode.jsRuby on RailsPythonFastAPIDockerPostgreSQLLLMsKYC/AMLFraud Detection
↓ scroll to explore
AI Systems
Multi-Agent · RAG · LLM
Payments & Risk
5 PSPs · KYC · Fraud
Full-Stack
Node.js · React · Prisma
Security
Network · Auth · IDS
01 — About
Who I Am
Operations and risk specialist with 3+ years at FundingPips — a global proprietary trading firm — owning the full lifecycle of payments, fraud detection, chargeback management, and KYC compliance across a 100+ country client base.
Engineered a cross-account card fraud detection system in Ruby on Rails that reduced high-risk account activations by 45%. Active practitioner in financial markets with hands-on experience in automated trading strategy development and risk-controlled execution.
Currently architecting and delivering a full-stack SaaS platform and pre-sales operation at SHAI For Artificial Intelligence in Amman. Combines deep compliance and risk discipline with hands-on engineering capability across the full development and automation stack. Concurrently building personal AI systems, trading bots, and multi-agent orchestration tools.
🎓
B.Sc. Software EngineeringAmman Arab University · 2018 – 2021
02 — Skills
Technical Stack
Full-Stack Engineering
AI Automation
Fintech & Risk
DevOps & Infrastructure
Security & Reliability
Languages & Certifications
03 — Experience
Work History
SHAI For Artificial Intelligence
Full-Stack Developer & Pre-Sales Manager
📍 Amman, Jordan · On-site · AI & Data Science Training Centre
May 2025 – Present
- Architected and built the company web platform from scratch: Node.js/Express backend, React 18 frontend with Vite and Tailwind CSS, containerised with Docker, deployed on Railway (nixpacks) with CI/CD via Vercel
- Built ShaiCRM from scratch — custom Prisma/SQLite data layer with Registration, User, Course, and support ticket models; lead lifecycle from intake through enrollment with React-based admin and CRM dashboard
- Engineered ShaiCom, a bilingual AI sales bot (Arabic/English) with intent classification and full conversation state machine; deployed as an embeddable widget across all company pages
- Built Discord.js v14 automation: on enrollment the system generates course-specific channel invites and dispatches onboarding notifications — zero manual intervention
- Implemented production security hardening: Helmet HTTP headers, JWT role-based auth, and layered rate limiting per route
- Serves as Pre-Sales Manager: B2B & B2C client communications, lead qualification, deal conversion, and sales pipeline management end-to-end
- Owns budget planning and financial analysis for operational projects, preparing structured cost models and tracking expenditure against forecasts
FundingPips
Project & Operations Specialist — Payments, Compliance & Risk
📍 Dubai, UAE · Hybrid · Global Proprietary Trading Firm
Mar 2022 – Dec 2024
- Engineered a cross-account card fraud detection system in Ruby on Rails — payment metadata, device fingerprinting, behavioural signals — reduced high-risk account activations by 45%
- Designed pre-activation risk screening framework (IP history, device fingerprinting, trading behaviour, payment data) — became baseline standard for all new account activations
- Managed end-to-end payment operations across 5 PSPs (Stripe, Checkout.com, MyFatoorah, NowPayments, Forumpay) across a 100+ country client base
- Investigated and resolved 15–25 chargeback and dispute cases daily; prepared evidence packages and coordinated with payment processors across the full dispute lifecycle
- Managed KYC verification and onboarding workflows for thousands of funded trading accounts, enforcing identity verification policies across a global client base
- Produced daily, weekly, monthly, and annual chargeback reports used by finance and operations teams for dispute ratio tracking and processor performance monitoring
- Implemented JIRA–Intercom integration and automated chatbot workflows, reducing manual ticket escalations and improving support response efficiency
- Authored and maintained 15+ SOPs covering onboarding, compliance review, payment handling, dispute management, and platform administration
- Managed operational workflows for a 50+ partner affiliate programme including KYC, contract preparation, and account provisioning
04 — Projects
What I've Built
RIVALUX
Concurrent LLM benchmarking engine. Runs the same prompt against multiple providers simultaneously, measuring latency, success rate, byte length, and token count. Scores each response with a local deterministic evaluator, stores runs in SQLite, and emits structured events to ORBCORE. Rich terminal dashboard included.
Explore Project →
Private Source · Case Study Available
OMNEX
Autonomous trading system requiring agreement from two independent layers before any trade executes: a multi-indicator technical analysis engine (multiple signals must align) plus an LLM reasoning layer that reads RAG-sourced news context and produces a confidence-weighted recommendation. Position sizing scales with signal confidence before order placement.
Explore Project →
Private Source · Case Study Available
SWARIX
Multi-agent orchestration system with a PlannerAgent that routes tasks by keyword to 7 specialised agents: RIVALUX for benchmarks, NetPulse for network, WraithNet for security response, BountyOps for bug bounty, CaseForge for case studies, PromptGuard for LLM security, and WebhookForge for payment webhooks. MemoryAgent and SynthesizerAgent run on every workflow.
Explore Project →
Private Source · Case Study Available
NETPULSE
Async network visibility collector using psutil to enumerate all active connections per snapshot window. Classifies every remote IP into six categories (loopback, private, link_local, multicast, reserved, public) at collection time. Host class distribution and anomaly counts feed WraithNet's security evaluation layer via a queryable FastAPI.
Explore Project →
Private Source · Case Study Available
WRAITHNET
Network security evaluator that learns a baseline from NetPulse snapshots and fires on two spike conditions: connection count exceeding 2.5× the learned average (floor: 20) and public destination count exceeding 3× the average (floor: 5). Risk scoring aggregates per-severity weights into a composite score that selects a response mode — observe, log, dry_run, or isolate.
Explore Project →
Private Source · Case Study Available
ORBCORE
Production orchestration server with SHA256 hash-chained NDJSON audit journals (jobs and audit are separate files). Four RBAC roles — admin, operator, auditor, viewer — with per-role rate limits (600 / 240 / 120 / 60 req/min). Atomic config writes via temp-file rename, log rotation at 5MB, bcrypt admin token, scoped per-worker auth, and optional TLS with client cert enforcement.
Explore Project →
Private Source · Case Study Available
SHAI For AI
Full-stack SaaS platform powering an AI training company — custom CRM, bilingual AI sales bot, Discord course-channel automation, and React/Node.js production stack. End-to-end sole engineer.
Explore Project →
Private Source · Case Study Available
Payments Fraud Detection Engine
Cross-account card fraud detection system in Ruby on Rails. Combines payment metadata, device fingerprinting, and behavioural signals at scale — reduced high-risk activations by 45%.
Explore Project →
Private Source · Case Study Available
KYC Sentinel
Automated identity verification and AML risk assessment platform. Combines document analysis, liveness detection, watchlist screening, and PEP checks into a unified risk-scored onboarding pipeline.
Explore Project →
Private Source · Case Study Available
ChargebackIQ
Dispute lifecycle management and chargeback intelligence platform. Tracks cases end-to-end, surfaces win-rate analytics, and provides processor trend reporting — built from 3 years of daily chargeback operations.
Explore Project →
Private Source · Case Study Available
WebhookForge
Local webhook testing and debugging tool. Captures, inspects, and replays webhook events from payment processors and third-party services — no external tunnelling required.
Explore Project →
Private Source · Case Study Available
PromptGuard
LLM security scanner running 50+ probe categories against target endpoints — secrets disclosure, system prompt leakage, role override, bypass attempts, schema extraction, and more. Detects API key patterns and secret-like strings in responses via regex, computes a composite vulnerability score (0–100), and integrates with BountyOps to auto-generate structured findings reports.
Explore Project →
Private Source · Case Study Available
05 — Network
Project Graph
▸ hover to identify · click any node to explore
06 — Blog
Notes & Thinking
AI Systems
When Two Systems Disagree: The Case for Hybrid Decision Layers
How ChimeraGuard and NetMon run a 10-rule deterministic engine and a local Ollama layer independently — and why the fallback to
rule-based-fallback when inference fails is a design choice, not a gap.Read Article →
Security
Baselining Behaviour: Network Anomaly Detection Without the Noise
How iMonitor and NetMon implement off / learning / active baseline states, why WraithNet uses 2.5× and 3× multipliers with absolute floors, and what NetPulse's 6-category IP classification changes about detection quality.
Read Article →
Engineering
Building Autonomous Systems That Audit Themselves
How OrbCore's SHA256 hash-chained NDJSON journals and Ruflo's memory_search / memory_store loop turn autonomous decisions into versioned, replayable, tamper-evident evidence — before the first action executes.
Read Article →
Recruiter Snapshot — Available Now
Full-Stack Developer with fintech operations experience. I build React, Node.js, Ruby on Rails, Python, and AI automation systems. My edge is rare: 3+ years of payments, KYC/AML, chargebacks, and fraud-risk operations — plus the engineering to automate them.
Open To
- Full-Stack Developer roles
- AI Automation Engineer roles
- Payments, Fraud & KYC/AML roles
Best-Fit Environments
- SaaS & fintech startups
- AI automation teams
- Payments infrastructure teams
- Risk & compliance technology
Work Model
- Remote · Hybrid · Global
- Location: Amman, Jordan
- Notice Period: 30 days
- Time Zone: UTC+3
Full Professional Resume
Resume — 3 Role Variants
Full work history, skills, and impact metrics. Tailored view for each target role: Full-Stack, AI Automation, or Payments Risk.
07 — Contact
Get In Touch
I'm open to full-time remote roles, freelance projects, and technical consulting. 30-day notice period. Available globally.
✅ Open to Work
Full-Stack Developer · Payments & Risk Operations · Fraud Analyst · Pre-Sales Manager
Remote · On-site · Hybrid | Global