Open to junior developer roles

I build the systems a business runs on.

Dispatch for a fibre-optic crew, sales forecasting for a telecom operator, coastline monitoring for a hotel group. Three final-year projects for real clients in Tunis, Tunisia — two of them running in production today.

See the work GitHub or run the terminal

Selected work

04 projects

Inbox Copilot

live

AI email assistant for Gmail.

Syncs your inbox in real time, sorts mail by category and priority, summarizes long threads, drafts replies in three tones, translates, schedules sends and chases follow-ups. Phishing detection puts SPF/DKIM/DMARC checks and link heuristics underneath the AI: the model can raise a verdict, never lower it. AI drafts never send without a human pressing send. Covered by 1,300+ automated tests. The live demo is limited to approved test users while Google verification is pending.

Next.js
TypeScript
Tailwind
Express
Prisma
PostgreSQL
Redis/BullMQ
Gemini API
Google OAuth
Vercel
Render
Neon
Upstash

Visit the live app Read the code

FibreConnect

live

Intervention management for a Tunisian fiber optic subcontractor.

A three-role platform (client, technician, supervisor) that coordinates fiber outages from first call to closing invoice. Zone-based dispatch, cash-payment collection with technician remittance tracking, and manual reassignment for zones with no coverage. Built for a real subcontractor, deployed in production.

Next.js 16
TypeScript
Prisma
PostgreSQL
NextAuth
Vercel

Visit the live app Read the code

TT KPI Dashboard

live

Sales performance + forecasting for a Tunisian telecom operator.

An interactive Streamlit dashboard with Prophet time-series forecasting projecting the remainder of the year, z-score anomaly detection flagging abnormal sales days, and regional breakdown per agency. Synthetic data mirrors the reporting structure of the Tunisian telecom regulator (INT).

Streamlit
Python
Prophet
pandas
Plotly
scipy

Visit the live app Read the code

Costalina

v1 complete

client: Iberostar

Citizen-science coastline monitoring for Iberostar.

A cross-platform Flutter app that turns beachgoers into contributors. Users photograph and geolocate coastal anomalies — erosion, pollution, debris — with real-time weather context and multi-language support (Arabic, French, English). Aggregated data gives researchers continuous, ground-truth monitoring at a fraction of the cost of official surveys.

Flutter
Dart
Node.js
Express
MongoDB
OpenStreetMap
Gemini

Read the code

How I work

Final-year computer science student at Collège LaSalle Tunis, graduating in 2026. Two of my three current projects were shaped by real client conversations, which has changed how I think about scope, edge cases, and what "done" actually means.

I care about production concerns early — real database, live deployment, seed data that mirrors real operations — because I've learned that retrofitting them later hurts more than doing them right the first time. When something breaks in production (and it has), I want to be the person who debugged it, not the one who never saw the logs.

Education: CS · Collège LaSalle Tunis · 2026
Location: Tunis, Tunisia
Languages: Arabic (native) · French · English
Availability: Open to junior roles · remote or on-site

Tools I have actually shipped with

Everything here appears in one of the projects above, not on a course syllabus.

Languages

Python
TypeScript
JavaScript
Dart
SQL

Frontend

Next.js
React
Flutter
Tailwind CSS

Backend & Data

Node.js
Express
Prisma
NextAuth
Streamlit
Prophet
pandas

Databases

PostgreSQL (Neon)
SQLite
MongoDB

Cloud & Tools

Vercel
Neon
Streamlit Cloud
OpenStreetMap
Open-Meteo
Git
GitHub Actions

Looking for a junior who has already shipped.

I am finishing my degree at Collège LaSalle Tunis and I am open to junior roles, remote or in Tunis, Tunisia. Pick whichever channel suits you — I answer the same day.