ML Financial Statement Analysis
Budget: $3,000 – $5,000 USD
We’re seeking a skilled freelancer to build a machine learning model that can extract, analyze, and assess the credit risk of businesses based on financial statements and bank statements in PDF format (English & Arabic).
This project involves data extraction, feature engineering, risk scoring, fraud detection, and model explainability. The end goal is a model that generates a risk score, raises red flags, and produces explainable insights from PDF data.
Key Objectives:
• Ingest scanned or digital PDFs of financial & bank statements
• Support Arabic and English documents
• Extract structured financial and behavioral features
• Perform credit risk scoring, trend analysis, and fraud detection
• Deliver a production-ready model and a scoring interface
Technical Requirements:
• OCR and parsing pipeline for Arabic + English PDFs (scanned and digital)
• Feature generation logic (financial ratios, bank behavior, trends, red flags)
• Supervised ML model to:
o Predict credit default risk
o Classify risk tiers
o Flag anomalies or fraud indicators
• Output layer/API that returns:
o Risk score
o Financial KPIs
o Red flags
o Model explanation (SHAP, LIME)
Expected Deliverables:
• Clean, modular ML model and codebase
• Feature engineering pipeline
• Sample datasets (or ability to work with ours)
• Scoring function or lightweight API
• Clear technical documentation
• Explainability report and visualizations
Ideal Freelancer:
• Experience in building ML models
• Strong in OCR or Extracting data, financial documents
• Understands financial ratios, cash flow behavior, and credit risk
• Highly preferred who has experience and understanding in Arabic text parsing
If you are interested, please share:
• Relevant portfolio to demonstrate your relevant experience
• Brief technical approach (document or Loom video)
• Tools/libraries you plan to use
• Estimated budget & delivery timeline
This project involves data extraction, feature engineering, risk scoring, fraud detection, and model explainability. The end goal is a model that generates a risk score, raises red flags, and produces explainable insights from PDF data.
Key Objectives:
• Ingest scanned or digital PDFs of financial & bank statements
• Support Arabic and English documents
• Extract structured financial and behavioral features
• Perform credit risk scoring, trend analysis, and fraud detection
• Deliver a production-ready model and a scoring interface
Technical Requirements:
• OCR and parsing pipeline for Arabic + English PDFs (scanned and digital)
• Feature generation logic (financial ratios, bank behavior, trends, red flags)
• Supervised ML model to:
o Predict credit default risk
o Classify risk tiers
o Flag anomalies or fraud indicators
• Output layer/API that returns:
o Risk score
o Financial KPIs
o Red flags
o Model explanation (SHAP, LIME)
Expected Deliverables:
• Clean, modular ML model and codebase
• Feature engineering pipeline
• Sample datasets (or ability to work with ours)
• Scoring function or lightweight API
• Clear technical documentation
• Explainability report and visualizations
Ideal Freelancer:
• Experience in building ML models
• Strong in OCR or Extracting data, financial documents
• Understands financial ratios, cash flow behavior, and credit risk
• Highly preferred who has experience and understanding in Arabic text parsing
If you are interested, please share:
• Relevant portfolio to demonstrate your relevant experience
• Brief technical approach (document or Loom video)
• Tools/libraries you plan to use
• Estimated budget & delivery timeline