AI & DATA SCIENCE · TANGIER, MOROCCO

Turning datainto insight.Building AI thathelps people work.

I’m Oussama Tabakh, a final-year AI & Data Science engineering student and freelance data analyst. I build dashboards, automation workflows, and applied AI solutions.

Explore selected work
Oussama Tabakh in a suit
Oussama TabakhOpen to a PFE & freelance projects
From data to action
01 / INPUT03 / OUTPUTDocumentsPDF / OCRDataCSV / SQLMessagesVOICE / TEXTAnalysis + AI02 / PROCESSINGUseful insightsTasks, handledCOMPLEX INPUTS. USEFUL OUTPUTS.
An illustration of how I work
Based in Tangier. Working across borders.RAG / OCR / ML / AUTOMATION

EXPERIENCE & EDUCATION

Bouygues Construction ITAI & backend internship · PFA completed
Smart Automation TechnologiesData Science & AI internship
FST · ENSIData Analysis degree · FST Tangier

01 / SELECTED WORK

From an idea
to something useful.

Applied AI, practical automation, and the engineering that brings them together.

01 / DOCUMENT INTELLIGENCE

Bouygues eDoc

Turning technical documents into structured information, with an OCR extraction pipeline built around the eDoc submission workflow.

  • Apache PDFBox
  • Azure AI Foundry
  • Java 17
  • Spring Boot
Explore case study — Bouygues eDoc
eDoc / Document intelligenceARCHITECTURE
eDocDOC—001
Technical document
NATIVE TEXT / IMAGE
Native textPDFBoxAzure LLM
Image / scanAzure VLMVisual reading
Structured fields
ReferenceDOC—001
CategoryTechnical
ValidationRules
DestinationeDoc

Extract → Classify → Validate → Pre-fill eDoc

Simplified diagram · Fictional document example
02 / RETRIEVAL-AUGMENTED GENERATION

University AI Assistant

A conversational interface for exploring documents about AI adoption in universities, powered by a PDF-to-RAG pipeline.

  • LangChain
  • Ollama
  • Chroma
  • Groq
Explore case study — University AI Assistant
University AI AssistantARCHITECTURE
01 / PDFIngestion

Documents · LangChain chunks

02 / OLLAMALocal index

Local embeddings · Chroma

03 / GROQGeneration

Question + retrieved context

A natural-language answer

Streamlit interface · Documents on AI in universities

Illustrative architecture · No simulated results
03 / MULTI-AGENT SYSTEMS

DevAgent

Four local agents coordinate specifications, development, testing, and deployment artifacts through shared project context.

  • Ollama
  • Python
  • Local LLMs
  • Agent orchestration
Explore case study — DevAgent
DevAgent / Agent workspaceINTERFACE
DevAgent
Screenshot of the DevAgent interface and its agent workspace.
Project screenshot · Actual interface
04 / MACHINE LEARNING WORKFLOWS

AI/ML Toolbox

A practical workspace for preparing datasets, exploring data, training models, evaluating results, and making predictions.

  • Python
  • scikit-learn
  • TensorFlow
  • Data analysis
Explore case study — AI/ML Toolbox
AI/ML ToolboxINTERFACE
AI/ML Toolbox
Screenshot of the AI/ML Toolbox application and its analysis tools.
Project screenshot · Actual interface
05 / VOICE & WORKFLOW AUTOMATIONLive

Medical Receptionist Agent

A voice receptionist used by a real medical practice to handle calls, manage appointments, and pass messages to the doctor.

  • ElevenLabs
  • n8n
  • Voice agents
  • Calendar workflows
Explore case study — Medical Receptionist Agent
Medical reception / Voice agentARCHITECTURE
Phone call → ConversationElevenLabs
n8n
Book
Reschedule
Cancel
Message the doctor
Illustrative workflow · No patient data

02 / HOW I CAN HELP

Your next challenge.
My kind of work.

For teams with data to understand, repetitive work to simplify, or an AI idea to put into practice.

01

Data into decisions

Make scattered data easier to understand and use, with analysis and reporting built around your questions.

  • Data preparation & exploratory analysis
  • Dashboards & visual reporting
  • Recurring reports & decision support
02

Workflows that work

Connect the repetitive parts of your day so your team can spend more time on the work that needs them.

  • n8n workflow automation
  • Voice agents & reception workflows
  • Calendar and application integrations
03

AI with a purpose

Build applied AI around a concrete task, from finding information in documents to serving a model in an application.

  • RAG assistants & document search
  • OCR & structured document extraction
  • Machine learning models & APIs

03 / THE PERSON BEHIND THE WORK

Curiosity,
put to work.

Oussama Tabakh in a suit
OUSSAMA TABAKH · TANGER

I’m finishing my engineering degree in AI & Data Science at ENSI, building on a degree in Data Analysis. Alongside my studies, I turn real problems into working tools through internships and freelance projects.

PFA completed. Seeking a six-month final-year internship (PFE), and available for freelance projects.

CV — English

Experience

  1. Current freelance work

    Data analyst & AI automation developer

    Independent

    Data analysis and applied AI projects, including automation workflows and a voice receptionist deployed in a medical practice.

  2. July – September 2026

    AI & Backend Development Intern · PFA

    Bouygues Construction IT · Casablanca

    Technical-document extraction, OCR processing, classification, validation, and integration into the eDoc submission workflow. PFA completed.

  3. April – June 2025

    Data Science & AI Intern

    Smart Automation Technologies · Tangier

    Built a Random Forest classification pipeline and a RAG assistant using LangChain and Chroma, with a Flask API and Streamlit interface.

Education

2025 – Present

Engineering studies in AI & Data Science · Final year

ENSI

2021 – 2025

Licence in Data Analysis

Faculty of Sciences and Techniques (FST)

Certifications & languages

  • DataCamp · Data Scientist Associate · 2025
  • DataCamp · Data Analyst Associate · 2025
  • EF SET · English C2
  • TCF · French B2
Loading GitHub activity…

05 / THE TOOLKIT

The right tools.
For the right problem.

A foundation in data and machine learning, with the engineering skills to make the solution usable.

01

Data & analytics

  • Python
  • SQL
  • Pandas
  • NumPy
  • Power BI
  • Tableau
  • Matplotlib
  • Seaborn
02

Applied AI & agents

  • LangChain
  • RAG
  • Ollama
  • Chroma
  • OCR
  • Azure AI Foundry
  • ElevenLabs
  • n8n
03

Machine learning

  • scikit-learn
  • PyTorch
  • TensorFlow
  • XGBoost
  • Feature engineering
  • Model evaluation
04

Engineering & delivery

  • FastAPI
  • Spring Boot
  • React
  • Next.js
  • PostgreSQL
  • Docker
  • GitHub Actions
  • ONNX

06 / LET’S TALK

Something in mind?
Let’s make it work.

A project to build, a data problem to untangle, or a PFE opportunity. I’d like to hear about it.

Oussama Tabakh — AI, Data & Automation