Ahmed Hasanin (Backend Tech Lead)

• Engineered an enterprise-grade Chatbot deployed at major financial institutions (NBK, Bank Misr, ABK). Integrated Azure Cognitive Services for Speech-to-Text (training models on domain-specific datasets/slang) and Text-to-Speech (tuned via SSML for natural interaction). • Led critical System Upgrades (Node.js v10 to v18, Socket.io 0.9 to v4, MongoDB v2 to v4), achieving a 10X performance boost while ensuring Zero Downtime for active banking customers. • Architected a solution for Arabic Sentiment Analysis by pipelining translation services with Opinion Mining. Replaced MongoDB text search with Elasticsearch to handle complex Arabic morphology and querying. • Designed a High Availability (HA) infrastructure using Docker to deploy replicas for MongoDB, Elasticsearch, and Redis, ensuring fault tolerance for real-time Socket.io connections. • Engineered robust infrastructure using EC2 and Load Balancers to handle peak traffic with policies for Autoscaling fully provisioned via AWS CloudFormation to ensure reproducible and automated deployments using Infrastructure as Code, while implementing end-to-end observability with CloudWatch (metrics), CloudTrail (auditing), and AWS X-Ray (distributed tracing) to reduce mean-time-to-resolution (MTTR). • Architected a scalable Serverless ecosystem using AWS Lambda, API Gateway, and DynamoDB, Integrated Amazon Cognito for secure authentication and ECS for container orchestration and Horizontal Scaling. • Secured the Serverless ecosystem by implementing OAuth2 workflows via AWS Cognito and JWT authorizers in API Gateway, enabling seamless Single Sign-On (SSO) for enterprise users.

• Contributed to a high-scale Question Bank system serving 500,000+ users across 31 universities using Node.js, Vue.js, MongoDB, and MariaDB, utilizing GitLab for version control and collaboration. • Implemented a Genetic Algorithm to automate exam generation based on complex criteria, reducing instructor workload by 70%. • Developed a resilient Online Exam System crucial during the pandemic, utilizing RabbitMQ for event queuing, Redis for caching, and Socket.io for real-time state management. • Refactored legacy code and optimized heavy Database Queries, reducing critical process execution time from 1 hour to 15 minutes and eliminating database locks. • Enhanced an Optical Mark Recognition (OMR) solution using C#, OpenCV, and NumPy for Image Processing, reducing bubble sheet grading defects by 20%