Product Engineer · AI-Powered Products · Technical Leadership
FadiThomas.
I build the parts of a product where the business logic is messiest.
7+ years of software engineering — full-stack product work across TypeScript/Node.js, React, Java and Python, from early-stage products to enterprise platforms — now at the intersection of product, AI and engineering.
I take features from discovery to delivery: framing the problem with stakeholders, scoping trade-offs, shaping the solution, and owning the build.
Recent work centres on AI-powered product features — LLM-based document understanding, agentic workflows, automation pipelines — and on connecting business intent to what actually ships.
Framing before building
Continuous discovery, PRDs and user flows, requirement framing, scoping and prioritisation, roadmap input, stakeholder alignment, metrics and KPI definition.
Models inside real products
LLM integration (Claude, vision models), AI agents, prompt engineering, document-understanding pipelines, AI-assisted development, n8n workflow automation.
Owning the whole loop
End-to-end feature ownership, technical direction, code review and mentoring, cross-team collaboration, working with non-technical stakeholders.
- Jul 2025 — PresentRemote
Product Engineer
Own AI-powered payroll and document features end-to-end, working directly with the executive team.
- Ran requirement conversations directly with the CEO, CFO and CX Director — translating business goals into scoped, sequenced work and pushing back where the trade-off didn't hold.
- Designed and built a document-understanding pipeline: pdf2pic rasterises incoming PDFs, Claude Vision extracts structured fields, BullMQ manages the job queue and retries, and a validation layer catches malformed output before it reaches downstream systems.
- Shipped a payroll interface with dual-source upload and field mapping, so operations can reconcile mismatched client formats without opening a ticket.
Removed a recurring manual data-entry step from operations and gave non-engineers a self-service path for data reconciliation.
TypeScript · tRPC · Prisma / Drizzle · PostgreSQL · Redis + BullMQ · React · Zod · Claude Vision · Jest
- 01
AI Document Processing Pipeline
Queue-backed service that rasterises incoming PDFs, extracts structured fields with Claude Vision, validates output, and hands clean data downstream. Built for a payroll context where formats are inconsistent and errors are expensive.
- 02
Raseedi — Smart Wallet Optimisation
Business-facing web application that matched advertisers to mobile users by budget classification.
- 03
Enrollet — Central Admission Service
A uni-assist-style central admission platform for Egyptian universities, simplifying applications and giving newer universities visibility they otherwise lacked. Graduation project.
- Languages
- TypeScript · JavaScript · Python · Java · SQL
- Frontend
- React · Next.js · Redux · Angular · Tailwind · React Hook Form · Zod
- Backend
- Node.js · NestJS · Express · tRPC · REST · GraphQL · Django · Spring Boot
- Data
- PostgreSQL · MySQL · MongoDB · Redis · Prisma · Drizzle · Tableau · ETL
- AI & Automation
- LLM integration · Claude Vision · AI agents · prompt engineering · n8n · BullMQ pipelines · AI-assisted development
- Infrastructure
- Docker · Kubernetes · AWS · CI/CD · Jenkins · Datadog
- Practices
- Agile · TDD/BDD · code review · Jira · Git · Figma
M.Sc. Global Software Development
Hochschule Fulda, GermanyB.Sc. Electronics & Computer Engineering
Nile University, Egypt — GPA 3.22 / 4.0Continuous Discovery Habits (Teresa Torres) · Android Development Nanodegree · Front End Development Track · IBM Internship (Blockchain) · BlueCloud Internship (Microsoft CRM) · 2nd Place Differential Equations & 2nd Place Linear Algebra, Nile University Research Forum
English — Advanced (IELTS C1 / 7.0) · German — A2 · Arabic — Native