Ziady MubaraqZiadyMubaraq

Architected an AI-powered SaaS platform for debt collection: TraceLink for automated skip tracing plus a multi-agent AI system autonomously running collection campaigns (intent detection, negotiation, follow-up).

Engineered LLM prompt strategies and a RAG pipeline grounding responses in live payment history and DPD risk data, reducing hallucination in production.

Orchestrated automation via n8n: CRON-triggered outreach combined with a periodic reply-monitoring loop, auto-routing ~8,000 daily debtor conversations to specialized agents by DPD risk, achieving a 40-60% reply rate.

Owned the full-stack across React/Vite and Node.js/TypeScript, backed by AWS infrastructure (EC2, S3, RDS, SQS, CloudWatch) with fault-tolerant, queue-based processing at scale.

Promoted to Lead Phase Instructor, mentoring a team of 6–8 instructors and 1,000+ students through project-based learning.

Delivered Full-Stack JavaScript and Front-End courses covering React, Node.js, Redux, React Native, GraphQL, TDD, Jest, and MongoDB.