Hammam Abdelmajed

Developed and optimized a LIDAR-based data visualization module for GIS, enabling measurement of road and building dimensions from interactive graphs; fine-tuned a Stable Diffusion-based image generation model for on-brand social content using LoRA and ControlNet; improved frontend performance of a large-scale web application; designed a company website with Three.js and custom GLTF 3D models; engineered backend to stream geospatial data from GeoServers via APIs for real-time rendering; built robust RESTful APIs for dynamic geospatial operations.

Built a low-latency WebSocket-based interface for a conversational AI querying blockchain data; integrated Apache Kafka for reliable streaming and backpressure; managed streaming AI results and conversation history with Zustand; deployed microservices on Kubernetes to enable scalable inference pipelines; addressed synchronization of token-by-token streaming with asynchronous UI updates to deliver a fluid typewriter effect.

Evaluated performance of real-time protocols for conversational AI; compared WebSocket and SSE implementations; concluded SSE reduced client-side complexity and improved reconnection handling; designed an AI-assisted content creation tool for sellers; used Apache Kafka as a streaming backbone to buffer and process AI requests; implemented optimistic UI updates with Redux Toolkit to mask latency for a seamless user experience.