Abdallah Wallyallah

Background

About Abdallah Wallyallah

I’m a Senior AI Engineer with 4+ years of experience building and deploying production-grade AI systems across industrial safety, real-time video analytics, and autonomous driving. I combine a strong research background with hands-on engineering, taking AI solutions from idea to scalable, monitored production.Currently, I design and deploy high-throughput, real-time computer vision systems handling 50+ concurrent camera streams, leveraging NVIDIA Triton, TensorRT, GStreamer, Docker, and Kubernetes. My work includes PPE compliance, slip & fall detection, fire and smoke detection, and multi-camera tracking in challenging industrial environments.Beyond vision, I’ve led the development of LLM-powered analytics agents that translate natural language into SQL, enabling non-technical stakeholders to extract real-time insights from production databases. I focus heavily on efficiency and reliability, using techniques like INT8 quantization, QAT, and system observability to reduce latency and operational cost.On the research side, I’ve contributed to peer-reviewed publications NILES), including GraphDrive, an interpretable Graph Attention Network framework for autonomous decision-making, and LiDAR–camera fusion models for end-to-end self-driving validated in CARLA simulations.Key Achievements & Experience • Deployed real-time AI systems supporting safety-critical industrial operations • Scaled inference pipelines to production workloads (50+ streams) • Delivered edge AI solutions using quantized models on constrained hardware • Published research bridging Explainable AI and autonomous systems • Built end-to-end pipelines spanning modeling, deployment, monitoring, and optimizationCore SkillsAI & GenAI: Computer Vision, LLMs, RAG, Agentic Workflows, GNNs, XAIEngineering: Python, FastAPI, Microservices, NVIDIA Triton, TensorRTMLOps & Deployment: Docker, Kubernetes, CI/CD, Model MonitoringDomains: Industrial Safety, Video Analytics, Autonomous DrivingMy focus is building AI systems that are scalable, interpretable, and dependable — especially where failure is not an option. I enjoy working at the intersection of research, engineering, and real-world impact.Always open to collaborations, technical discussions, and challenging problems at the frontier of AI.