[nom]

AI Engineer

Meknes, Morocco — [téléphone] — [e-mail]

[lien] — [lien]

Languages: Arabic (Native), French (Advanced), English (Fluent)

Professional Summary

AI Engineer specializing in LLM agents, RAG systems, and production computer vision. Currently building local-first AI automation at Segula Technologies with Ollama multi-agent councils, explainable anomaly detection, and full-stack deployment

(FastAPI, Next.js, Docker). Experienced shipping air-gapped GenAI pipelines, real-time CV inference, and MLOps-ready model serving.

Technical Skills

ˆ Generative AI & LLMs: RAG, Ollama, LangChain, multi-agent orchestration, LoRA fine-tuning, Self-RAG, DeepEval.

ˆ Computer Vision: PyTorch, UNet, YOLO (Ultralytics), OpenCV, real-time inference, Albumentations.

ˆ MLOps & Serving: FastAPI, Docker, MLflow, Airflow, GitHub Actions, SHAP, drift detection.

ˆ Stack: Python, Hugging Face, Qdrant, ChromaDB, Celery, PostgreSQL, React/Next.js.

Professional Experience

ˆ · Segula Technologies · Casablanca, Morocco

Final-Year Intern (PFE) — AI Engineer · Feb 2026 – Present

• Built AutoCrashCheck, a local-first audit platform combining a 5-agent Ollama LLM council, Isolation Forest, and SHAP explainability for crash-simulation decks.
• Delivered Next.js, FastAPI, PyQt, and CLI surfaces from one shared AuditService for OEM-confidential, air-gapped workflows.
• Containerized the stack with Docker Compose and optional Airflow retraining DAGs; 100% on-prem LLM inference.
• Cut manual engineering review time by 70% via automated feature extraction and JSON/PDF risk reports.

ˆ GEOFLY · Meknes, Morocco

Computer Vision Engineer Intern · Jul 2025 – Sep 2025

• Shipped a production Light UNet reaching 83.34% IoU at 312 FPS GPU inference from aerial imagery.
• Deployed FastAPI model serving with Docker Compose, Redis caching, and Prometheus/Grafana monitoring.
• Built GIS polyline extraction with automated GeoJSON export for downstream geospatial applications.

ˆ Saiket System · Remote / Meknes

ML Engineer Intern · Mar 2025 – May 2025

• Deployed churn prediction models (80% recall, 70% precision) through containerized CI/CD pipelines.
• Managed experiment tracking and model registry in MLflow for reproducible promotion to production.
• Automated training-data pipelines with GitHub Actions and Docker.

Key Technical Projects

ˆ Sovereign Enterprise RAG | Ollama, Qdrant, FastAPI, Celery, MLflow, React

Air-gapped RAG platform with hybrid retrieval, async Celery ingestion, Self-RAG verification, and DeepEval A/B evaluation targeting 35% hallucination reduction.

ˆ Building Segmentation AI | PyTorch, UNet, FastAPI, Docker

End-to-end CV pipeline for automated building segmentation from aerial imagery with production API deployment and GIS-ready outputs.

ˆ Scholarship Pathway Chatbot | LangChain, Ollama, Streamlit, Python

Local RAG chatbot over structured scholarship data using Mistral and embedding retrieval for student pathway guidance.

Education

ˆ · ENSAM Meknes · Meknes, Morocco

State Engineering Degree — Industrial Engineering, AI & Data Science · Sep 2021 – Present

Certifications

ˆ AWS Cloud Practitioner Essentials – AWS Training (Jan 2026)

ˆ Six Sigma Yellow Belt – ECLEE (Apr 2025)

ˆ Python Machine Learning & Data Science Bootcamp – Udemy (Jan 2025)

1