I build systems
that act —
and prove what they did.
I’m Malek El Badawi El Najjar, a Computer Engineering student and independent developer building AI automation, agent runtimes, API integrations, and full-stack products. I work across n8n, TypeScript / Next.js, Python / FastAPI, PostgreSQL, OpenAI, and Anthropic — from messy business process to deployed, verifiable system.
THREE SYSTEMS · THREE DIFFERENT TRUST BOUNDARIES
Different problems. Different constraints.
Same rule: verify the real result.
AGENTIC SOFTWARE ENGINEERING
CodePilot
A coding-agent runtime where the model can propose a change — but cannot declare its own work correct.
Authority boundary. Validation commands come from trusted workspace configuration, not model-controlled arguments.
Repair with limits. File, line, byte, and repair-attempt budgets are checked before mutation.
Evidence-backed completion. Verification is persisted only when the exact validated workspace and config still match.
62c7e3d19a0b…9af1
✓ VERIFIED
AI AUTOMATION / BUSINESS OPERATIONS
AutoFlow
A lead operations platform where n8n coordinates intake, AI scoring, authenticated write-back, and outbound notifications.
Real orchestration. A webhook enters n8n, gets validated, scored through structured AI output, written back through an authenticated internal API, then routed to notifications.
Retry-safe behavior. Notification paths are idempotent so a workflow retry does not duplicate outbound alerts.
Public demo isolation. The browser demo is synthetic and read-only, with no path to production data or side effects.
AI OPERATIONS CONTROL PLANE
OpsPilot
A control plane where AI may propose business actions while authorization, approval, execution, verification, and audit remain outside model control.
Five-step gate. Tool registration, agent visibility, integration state, workspace policy, and human approval are checked before execution.
Cross-integration runs. One task can search CRM, read mail, draft a response, pause for approval, and schedule a meeting.
Verified writes. External writes are checked against a fresh provider read and recorded in an append-only audit trail.
Not everything I build is an AI lab.
I also work directly with businesses: understand the problem, design the solution, ship it, and keep it usable after launch.
THE PERSON BEHIND THE SYSTEMS
I’ve been taking computers apart — physically or logically — for as long as I can remember.
I’m currently studying Computer Engineering at Beirut Arab University. Outside university, I build software because I care about what happens after the demo: where data lives, what breaks, who has authority, how retries behave, and whether the system can prove what it actually did.
Before freelancing, I worked customer-facing jobs in finance, technology retail, and hospitality. That background is useful now: I think about the person operating the system, not only the code behind it.
06 / CONTACT