Oussama Lakrafi

Architected a scalable microservices healthcare platform enabling real-time video consultations and AI-driven chatbot support. Implemented secure doctor verification, automated prescription generation with QR codes, and centralized authentication via Keycloak. Integrated Apache Kafka for event-driven notifications and RNN models to automate toxic feedback detection and content moderation.

Provisioned a production-ready AWS EKS cluster using Terraform for Infrastructure as Code. Automated a full CI/CD pipeline featuring Jenkins for CI, SonarQube for static code analysis, and ArgoCD for GitOps-based CD. Integrated a complete monitoring stack using Prometheus and Grafana to ensure high observability and scalability of the three-tier application.

Designed a stateless microservices architecture using Kafka as a central event bus for real-time inter-service communication. Implemented specialized services for ML inference, batching, and alerting to process live patient vitals with sub-second latency. Optimized the stack using Redis for alert deduplication and built an analytics lakehouse with S3, Glue, and Athena, surfaced through a Next.js WebSocket dashboard.

Developed a full-stack auction platform for real-time bidding. Implemented JWT authentication, WebSocket communication with STOMP.js, live notifications, and interactive charts for bid trend analysis.

Built a full-stack system with role-based JWT authentication. Features include employee/task management, department assignment, WebRTC for video calls, and Socket.IO for instant messaging.

End-to-end containerized data streaming pipeline using Airflow for data ingestion to Kafka, Spark Structured Streaming for processing, and Cassandra for storage. Monitored via custom dashboards.

Developed a real-time IoT platform to remotely operate a smart car via a Next.js dashboard and AWS. Integrated WebRTC and Kinesis for low-latency video streaming, while managing car movement through IoT Core and MQTT. Implemented an EC2-hosted TensorFlow model for real-time object detection with live visualization and DynamoDB logging.

Built a fully observable AWS infrastructure using CloudFormation with public/private subnets, NAT Gateway, and security groups. Deployed a FastAPI application behind an ALB with ASG. Configured CloudWatch logs/metrics, AWS X-Ray distributed tracing, and performed load testing with k6.

End-to-end ELT data pipeline orchestrated with Airflow. Leveraged Astronomer Cosmos to render dbt models dynamically as Airflow DAG tasks. Snowflake serves as the cloud data warehouse with staging, intermediate, and marts layers. Monitored data quality via dbt tests inside Docker constraints.