Contact
Bring your concepts
to the next LEVEL

Let's connect
to scale your IDEAS
Interactive lamp
Taj Eddine Marmoul
CVContact

Who I am

Taj Eddine Marmoul.

Junior AI & Full-Stack Engineer focused on agentic systems, RAG, and document AI.

Graduating soon as an AI and computer science engineer, currently building an MCP-based agentic platform at CNRS LAAS in Toulouse. My work sits between applied AI, backend systems, and product interfaces.

Toulouse / CasablancaAvailable Oct 2026Remote / relocation

Selected work

01

Agentic Web Platform

CNRS LAAS internship / 2026

Role: AI Platform Engineer Intern

Tools

FastMCPFastAPIReactMongoDBDocker
Problem
Researchers and lab teams need internal apps, forms, and data workflows, but turning each request into a working database, API, and interface is slow when every piece has to be coded manually.
Built
Built an agentic web platform where users describe what they need in natural language, then agents execute controlled MCP tool calls to generate database structures, backend logic, and UI pieces.
Evidence
Deployed in a lab environment used by about 200 daily users, supporting 8,000+ generated forms and 3,000+ created apps.

02

L'Cachet

Independent AI product / 2024-present

Role: Full-Stack AI Engineer

Tools

Next.jsTypeScriptPostgreSQLPrismapgvectorGemini
Problem
Moroccan legal and administrative information is scattered across long documents, PDFs, and official procedures, which makes it hard for citizens and professionals to find grounded answers quickly.
Built
Built a multilingual RAG platform that ingests official documents, indexes them for vector search, and answers questions with source-aware retrieval instead of generic chatbot responses.
Evidence
Processed 6,600+ legal documents and 2,500+ official procedures into 85,000+ embeddings with authentication, usage controls, shared sessions, and four-language support.

03

PDF2Text-Arabic

Document AI engine / 2024-present

Role: Python / Document AI Engineer

Tools

PythonPyMuPDF
Problem
Arabic legal PDFs often contain tables, footers, OCR noise, and right-to-left layout issues that break normal text extraction and weaken retrieval quality.
Built
Built the extraction engine behind L'Cachet's ingestion pipeline, handling parsing, cleanup, OCR fallback, table extraction, and RTL text reconstruction.
Evidence
Improves the quality of legal-document ingestion before chunking, embedding, and retrieval, so the RAG system receives cleaner source text.

Work history

Recent work.

A quick read on the systems, dashboards, games, and computer-vision work I shipped across internships and remote roles.

Feb-Aug 2026

AI Platform Engineer Intern

CNRS LAAS

Built an internal platform where researchers describe the app they need, then agents create the database structure, backend logic, and interface through controlled tool calls.

Designed for lab workflows with approval, auditability, and safer generation than raw code output.

Jul-Aug 2025

BI, Data & DevOps Intern

Royal Air Maroc

Built a flight-operations dashboard for operational teams to monitor data and KPIs from one application instead of relying on slow manual reporting workflows.

Reduced dashboard load time by 60%, automated ETL/KPI reporting, and helped deploy web/local versions used inside Royal Air Maroc systems.

Oct 2024-Jul 2025

Unity Game Developer

Studyo.io

Built educational mini-games that turn lesson concepts into interactive exercises and track how students engage with them.

Delivered 7 Unity games and used analytics feedback to improve student engagement metrics.

Jul 2024

AI & Full-Stack Intern

1337 Coding School

Led a small team building a real-time security system that detects people across cameras and verifies identity through face recognition.

Combined computer vision models with a WebSocket app, role-based access, and multi-camera monitoring.

Capabilities

What I work with.

The tools, systems, and habits I use to build AI products and keep improving outside project work.

Capabilities

Core stack.

AI / Agents

MCPFastMCPRAGtool callingvector searchLLM APIs

Backend / Data

PythonFastAPIPostgreSQLMongoDBRedisPrismaREST APIs

Frontend / Infra

ReactNext.jsTypeScriptDockerNginxLinuxTLS

Outside work

What keeps me sharp.

Building side projects

I like turning rough ideas into small working products, especially around AI tools, automation, and document workflows.

Games and interactive systems

Game development is where I learned to care about feedback loops, performance, and how people feel while using software.

Reading about AI and systems

I follow practical AI engineering, backend architecture, and product design so my work stays grounded in what can actually ship.

Availability

Contact.

Available from October 2026 for AI engineering, RAG, agentic systems, and full-stack AI product roles.

Start a conversation
LinkedInGitHub