•
Designed and developed scalable machine learning models for fraud prevention and personalization, fine-tuning large language models (e.g., GPT-3.5, BERT, RoBERTa) for biometric security applications, enhancing face matching, liveness detection, and spoof prevention systems, achieving a 25% improvement in detection accuracy and meeting performance and latency standards
•
Implemented biometric validation systems for real-time decision-making, including face liveness detection for user authenticity verification, face matching for identity confirmation, and OCR-based data extraction for automated document processing, improving system reliability by 20% and applicable to iGaming fraud prevention and user engagement
•
Led a Know Your Business (KYB) compliance automation project by implementing transformer-based NLP models, leveraging embeddings (e.g., Sentence-BERT) and retrieval-augmented generation (RAG) techniques on large-scale datasets, reducing verification time by 30% and improving compliance accuracy by 20% while ensuring ethical AI practices
•
Optimized LLM prompts, tokenization strategies, and ML pipelines for real-time document analysis, improving token efficiency by 15% and enabling faster decision-making workflows in high-stakes compliance environments, with integration into production systems
•
Designed and deployed scalable AI/ML solutions on AWS (EC2, S3, Lambda, SageMaker), utilizing MLOps practices (Docker, Kubernetes) to enhance system performance, achieving 99.9% uptime, reducing latency by 10%, and supporting model versioning, testing, deployment, and monitoring
•
Collaborated with cross-functional teams, including product managers and stakeholders, to integrate NLP and biometric pipelines with existing security systems, translating business needs into AI-driven solutions, streamlining data processing, and improving operational reliability
•
Built and integrated RESTful APIs for serving models in real-time environments, enabling seamless production integration and compliance with industry standards
•
Reference: Hazem Abdullah (CTO), +90 535 880 6432