Data Scientist & AI Engineer · Agentic AI & LLM Systems · INPT '26

LinkedIn · Email · GitHub

I design and deploy production-grade AI systems that turn complex data into actionable business insights, enabling non-technical teams to make data-driven decisions at scale.
Specialized in LLMs and multi-agent architectures, I build platforms that reduce time-to-insight from days to minutes while significantly lowering operational costs.

AI / ML
PyTorch TensorFlow Scikit-learn Keras HuggingFace OpenAI LangChain LangGraph Google ADK LlamaIndex MLflow Pandas NumPy SciPy NLP (NLTK) OpenCV Ragas n8n Dataiku

Data Engineering
Spark Hadoop Airflow pySpark SQL PostgreSQL MongoDB Talend DAGs Data Warehousing

Backend & Infra
FastAPI Flask Django Spring Boot REST APIs Docker Git CI/CD React Angular

Languages
Python R Java C/C++ Bash

Development of an agentic platform enabling business users to query databases and generate visualizations through natural language, without SQL. Designed an architecture orchestrated by an AI agent to specialized agents (Text-to-SQL, Data Visualization, Data Cleaning, Feature Engineering), with schema-based dynamic SQL generation, secure sandbox execution, and multi-turn conversational memory.

Impact: Reduced dependency on data teams by 85%, cut time-to-insight from days to minutes, increased adoption of data tools across business teams, and significantly lowered data extraction and analysis costs.

Dataiku Python LangGraph FastAPI Spark React Spark Hive OracleDB OpenMetadata

Designed a data platform to transform Google Maps customer reviews into actionable insights for telecom competitive benchmarking in Morocco. Built an industrialized medallion data pipeline with LLM-based enrichment (sentiment & drivers: network, agency, customer service, FTTH, etc.) and structured it into a data warehouse for multi-dimensional analysis through executive dashboards.
Impact: Analyzed 100k+ reviews with time-to-insight under 1 hour, identified 15% more critical pain points MoM and YoY, improved competitive benchmark granularity by 25%, and directly supported strategic decisions (service quality, investment prioritization). Python pySpark Api/Selenium SQL Power BI Dataiku

Multi-agent AI platform automating marketing content generation and competitive analysis, improving speed and strategic responsiveness for business teams by designing a multi-agent architecture (LangGraph) combining multimodal agents (vision) for visual content creation and research agents for telecom offer benchmarking, with orchestration, LLM-based synthesis, and structured deliverable generation for business teams.
Impact: Reduced marketing production time by 70%, enabled automated generation of dozens of visuals per week, cut competitive intelligence time by 3x, and improved strategic responsiveness to market offers.

Python Flask LangGraph LLMs LiteLLM Tivaly Pydantic

• Working on: Advanced multi-agent AI systems for enterprise data analytics
• Exploring: Agentic workflows, LLM optimization, and scalable AI architectures
• Open to: AI Engineering roles, Data Science positions, and impactful AI projects