Ismail Allouch

Who I Am

Engineering intelligence
into real-world systems.

I'm a 4th-year engineering student at ENSAM Meknès, specializing in Software Engineering & Intelligent Systems. My journey in AI started with classical machine learning and evolved into a deep passion for autonomous agent architectures.

Today, I focus on building systems where AI agents reason about problems, retrieve relevant knowledge, and take autonomous action — from multi-agent RAG pipelines with LangGraph and CrewAI to production automation workflows with n8n.

I believe the most impactful AI isn't just a model — it's an orchestrated system that connects LLMs with tools, data, and human oversight. That's what I build.

🤖
Agentic AI
Multi-agent systems, LangGraph, CrewAI, MCP & A2A protocols

🔍
RAG Architectures
Self-RAG, Agentic RAG, vector search, contextual retrieval

⚡
AI Automation
n8n workflows, LLM orchestration, event-driven pipelines

🏗️
Full-Stack Dev
React, Node.js, Spring Boot, PostgreSQL, Docker

Experience

Where I've built
and competed.

Jan — Mar 2026

AI Automation Engineer (Alternance)

M'covery

Designed and deployed AI-powered automation workflows using n8n for business process optimization. Integrated LLM-based pipelines for intelligent data processing and decision-making automation.

n8n LLM Automation AI Pipelines

Jul 2025

Full-Stack Developer (Internship)

Eventi Production

Developed a MERN-stack client review application with thematic analysis using NLP and LDA topic modeling. Implemented web scraping from Booking & Google Maps with API integration.

NLP LDA MERN Scraping

2025

Semi-Finalist & Team Leader — Techathon 2025

Capgemini Engineering Morocco × ENSAM Network

Led a team to the semi-finals of Capgemini's 5th AI innovation challenge — "Smarter Worlds with AI: Think Big. Build Smart. Shape the Future." Designed an AI-driven solution addressing environmental and societal challenges.

Leadership Innovation AI

2025

Winner — n8n AI Automation Competition

Club IA · ENSAM Meknès

Won first place in an AI automation competition focused on building intelligent workflows with n8n.

1st Place n8n

Projects

Things I've built.

From multi-agent AI systems to full-stack platforms — each project solves a real problem with thoughtful engineering.

🧠

Agentic AI

Multi-Agent RAG System for Document Intelligence

Built a multi-agent system leveraging Agentic RAG with specialized agents for document retrieval, query routing, answer generation, and self-reflective validation (Self-RAG). Orchestrated via LangGraph with dynamic tool calling and MCP protocol integration. Each agent has a defined role and communicates through a stateful graph, enabling complex reasoning chains over large document corpora.

LangGraph LangChain CrewAI Self-RAG MCP Protocol Python

                        // Agent Architecture

                        graph = StateGraph()

                        graph.add_node("retriever")

                        graph.add_node("router")

                        graph.add_node("generator")

                        graph.add_node("validator")

                        // Self-reflective loop

                        graph.add_edge(

                          "validator" → "retriever"

                        )

🧭

Full-Stack AI

JobGPS — AI Career Guidance Platform

Microservices platform (3 services) for AI-powered career guidance. RAG-based system with ChromaDB & Sentence Transformers for intelligent job matching, CV analysis, and personalized cover letter generation. Knowledge base: 50+ IT jobs, 200+ classified skills.

React 19 Express.js Flask Gemini ChromaDB RAG

🏫

Platform

2D-Campus — University Housing Platform

Full-stack platform for Euromed Fès with RAG chatbot (LangChain + Gemini + pgvector), JWT/bcrypt security, interactive SVG campus map, analytics dashboards and microservices. Agile/Scrum methodology.

React Node.js Supabase LangChain pgvector

🌱

IoT × AI

Agri-Gen — No-Code SaaS for Smart Agriculture

Cloud platform transforming natural language into IoT firmware via generative AI (Gemini 2.5 Flash). 96% reduction in IoT configuration time. Automated sensor/actuator code generation with Jinja2 templates and real-time simulation.

React Flask Gemini AI Jinja2 IoT

📊

Automation

Intelligent CRM & Campaign Orchestration

Event-driven architecture with two automated pipelines (Marketing & Customer Service). LLM-powered content generation, sentiment analysis, intent detection, and human-in-the-loop escalation via WhatsApp & Gmail APIs.

n8n Supabase Gemini API WhatsApp API

📸

Creative AI

Social Media Content Automation

Automated pipeline for AI-generated image and video creation with scheduled posting to Instagram and Facebook via Meta API integration. Fully autonomous content lifecycle.

n8n AI Generation Meta API

Tech Stack

Tools of the trade.

🤖

LLM & Agentic AI

LangChain LangGraph CrewAI RAG Self-RAG Agentic RAG MCP Protocol A2A Protocol

📈

ML / DL / NLP

scikit-learn TensorFlow PyTorch NLP Deep Learning Reinforcement Learning

⚡

AI Automation

n8n Workflow Orchestration Webhooks API Integration

🔧

Backend & DevOps

Node.js Express Spring Boot ASP.NET Flask REST API Docker Web Services

🎨

Frontend

React.js TailwindCSS HTML/CSS JavaScript

🗄️

Databases

PostgreSQL SQL NoSQL Supabase pgvector Sharding Partitioning Indexing

Programming Languages

Python Java C++ JavaScript SQL R

Ismail
Allouch

Software & AI Engineering Student building
autonomous systems that reason, retrieve & act.

Engineering intelligence
into real-world systems.

Where I've built
and competed.

Things I've built.

Tools of the trade.

Always learning.

Let's build something
extraordinary.

Software & AI Engineering Student buildingautonomous systems that reason, retrieve & act.

Engineering intelligenceinto real-world systems.

Where I've builtand competed.

Things I've built.

Tools of the trade.

Always learning.

Let's build somethingextraordinary.

Ismail
Allouch

Software & AI Engineering Student building
autonomous systems that reason, retrieve & act.

Engineering intelligence
into real-world systems.

Where I've built
and competed.

Let's build something
extraordinary.