بهاء كعكي

Data Scientist في CME

• لبنان
• أشغل هذه الوظيفة منذ يناير 2021

Developed and deployed a Language Model (LLM) that led to a 20% increase in user satisfaction by providing quick
and accurate responses to electronic invoicing inquiries. The model also reduced user support response times by 30%.
• Engineered a robust bank transaction classifier model with an 80% accuracy rate, saving 50 hours of manual data entry
monthly and enabling more accurate financial analysis.
• Developed and implemented a cutting-edge similarity search model for Small and Midsize Businesses, delivering
actionable recommendations that boosted client performance by 10%.
• Automated the accounting transaction classification process by introducing a Random Forest Classifier model. This
new model decreased processing time by 40%, resulting in significant cost savings.
• Successfully deployed and maintained models with a 99% uptime, collaborated with cross-functional teams, and
reduced operational costs by 10%.
• Optimized an OCR Quality of Extractions model. This streamlined the invoice review process for users and reduced
manual data correction efforts by 30%.
• Implemented comprehensive monitoring functions, resulting in a 20% reduction in model-related incidents, and
provided valuable insights to stakeholders, contributing to a 15% increase in stakeholder satisfaction.