Get in touch about AI engineering and consulting — RAG systems, computer vision, production ML. Based in Rabat, Morocco; open to remote and contract work.
AI engineer building production medical AI — retrieval-augmented clinical assistants and computer-vision triage pipelines that ship behind real safety constraints.
Details
Email: adil.eddarif@gmail.com
Phone: +212 605695329
Location: Rabat, Morocco (Africa/Casablanca)
Availability: Available for AI engineering and consulting work
Open to: Remote, Hybrid (Rabat / Casablanca), Contract, Full-time
What to get in touch about
Adil Eddarif is an AI engineer based in Rabat, Morocco, building production medical AI at the Cheikh Zaïd Foundation. He built a clinical RAG assistant whose emergency-case recall he raised from 50% to 100% by catching a safety-routing bug with an LLM-as-judge evaluation harness, and a two-stage chest X-ray triage pipeline (ResNet50 + YOLOv8) reaching 96.93% accuracy and 99.71% AUC on a 522-image held-out test set.
AI engineer Morocco
AI engineer in Morocco
hire AI engineer Morocco
medical AI engineer Morocco
AI engineer North Africa
AI engineer Maghreb
machine learning engineer Morocco
LLM engineer Morocco
RAG engineer Morocco
RAG evaluation engineer
healthcare AI engineer
medical imaging AI engineer
computer vision engineer Morocco
satellite imagery AI engineer
geospatial computer vision engineer
remote sensing deep learning
satellite object detection GeoJSON
MLOps engineer Morocco
AI consultant Morocco
freelance AI engineer Morocco
remote AI engineer Africa
AI engineer Rabat
AI engineer Casablanca
chest X-ray AI screening
Recent work worth asking about
Raised emergency-case recall from 50% to 100% on a clinical RAG assistant by finding a safety-routing bug with an LLM-as-judge evaluation harness.
Shipped a two-stage chest X-ray triage pipeline (ResNet50 + YOLOv8) at 96.93% accuracy and 99.71% AUC on a 522-image held-out test set.
Built a 1,000-question golden evaluation dataset with independent faithfulness and hallucination scoring.
Replaced heuristic financial forecasting rules at Attijariwafa Bank with supervised models at 0.90 R², tracked end-to-end in MLflow.
Delivered a satellite computer-vision pipeline for GeoAP across 10 object classes — YOLO11m detection at 0.836 mAP50 and a fine-tuned SAM1 ViT-H segmenter at 0.804 mIoU — exported as GIS-ready GeoJSON.
Cut SAM1 ViT-H fine-tuning from 88 minutes to 5 per epoch on a single T4 GPU by pre-computing and caching image embeddings across 9,958 tiles.
Reached >99% recognition accuracy on Tifinagh script OCR across a 25,740-image dataset.
Before you write
The pages below cover most of what people ask first — the measured outcomes, the stack, and the reasoning behind the systems.