ACHRAF CHAHBOUNE

Document AI Engineer

French & Arabic Document Extraction

Data Engineer - Kafka, PySpark, Airflow

PhD Candidate - Multi-Agent RL

ACHRAF CHAHBOUNE

Document AI Engineer

French & Arabic Document Extraction

Data Engineer - Kafka, PySpark, Airflow

PhD Candidate - Multi-Agent RL

Invoice Automation & Document Intelligence

  • Created By: Achraf Chahboune
  • Date: 2025
  • Type: Production Platform
  • Stack: LLM, FastAPI, PostgreSQL, GCP

Finance teams lose hours to manual invoice and bank-statement processing, and every manual step is a chance for a costly error. This platform automates the entire flow: it ingests financial PDFs, extracts the structured data with LLM-based document intelligence, and reconciles every figure against business and legal rules before producing compliant, filing-ready documents.

It is built as a clean microservice architecture – FastAPI services on PostgreSQL handle extraction, validation, and orchestration, with a React and TypeScript front end and secure multi-user access. A continuous LoRA/PEFT fine-tuning loop keeps extraction accuracy improving as more documents flow through the system.

The platform runs in production on GCP Cloud Run with Docker, automated CI/CD, and end-to-end monitoring and observability – an auditable, hands-off pipeline that turns a slow back-office task into a reliable automated service.