SIGNAL//OS
Yassine El Amrani
AI & Computer Vision Engineer
Building supervised AI systems where every output is traceable back to evidence.
Fast, print-friendly scan: skills, experience, projects, contact.
Signals
Invoice Intelligence
User approvedMultilingual invoice extraction for Moroccan SMEs
Problem
Small Moroccan accounting teams manually re-typed invoices in French, Arabic and mixed-language formats, losing hours per week.
Approach
Built an OCR + layout-aware extraction pipeline with a human review queue: every low-confidence field is flagged for approval before it reaches the accounting export.
Outcome
Cut manual entry time by an estimated 70% in pilot testing with three small accounting teams.
Pipeline evidence
Projects
Line Inspector
Computer Vision
Industrial defect detection on production lines
A CV inspection prototype that flags surface defects in real time and routes uncertain cases to a human reviewer.
Multilingual Doc Extractor
NLP / OCR
General-purpose document extraction engine
Generalizes the invoice pipeline's layout-aware extraction to contracts, ID documents and forms across three languages.
Model Supervision Dashboard
MLOps
Human-in-the-loop review tooling
An internal tool for reviewing, approving or rejecting model predictions before they reach production, with full audit history.
Experience
Machine Learning Engineer
2024 — PresentAtlas Vision Labs (demo)
Owns the computer-vision inspection pipeline from data labeling to production monitoring.
Freelance Computer Vision Consultant
2022 — 2024Independent (demo)
Delivered document-intelligence and inspection prototypes for small manufacturing and fintech teams.
Education
Engineering Degree, Artificial Intelligence & Data
National School of Applied Sciences (demo) · 2019 — 2022
Focused on applied machine learning, computer vision and distributed data systems.
Certificates
Open to AI engineering roles and computer-vision contract work
Contact for AI/CV roles