Hello
I'm Abdullah Amawi
ML Engineer · Infrastructure · GPU Data Pipelines · HPC Systems
I specialise in ML infrastructure and HPC systems, designing and optimising GPU-accelerated data pipelines for large-scale machine learning workloads. My M.Sc. thesis, completed in collaboration with GWDG, benchmarked NVIDIA DALI, TFRecords, and WebDataset across PyTorch and TensorFlow on SLURM-managed clusters. Currently working independently on ML infrastructure consulting, GDPR-compliant self-hosted infrastructure, and AI automation. Based in Germany.
Projects
Optimizing I/O Performance of Scalable ML Workflows in HPC Systems
This thesis investigates data-loading bottlenecks in large-scale deep learning workflows. By integrating GPU-accelerated pipelines such as NVIDIA DALI, experiments show significant reductions in training time and improved GPU utilization.
TensorRigs.com
AI Hardware & ML GPUs resource for Deep Learning, PyTorch, TensorFlow & Neural Networks. Expert guides for machine learning GPUs (NVIDIA RTX, A100, H100), deep learning systems, and AI workstations.
TensorDocs - HPC & ML Documentation
Comprehensive documentation website built with Starlight (Astro) covering high-performance computing, deep learning workflows, and ML optimization techniques.
Starlight Documentation for Large-Scale Machine Learning in HPC
Additional documentation from thesis work addressing software environment issues and challenges in HPC and deep learning workloads.
Sentiment Analysis on Course Feedback
NLP project exploring sentiment analysis on student feedback from a teaching seminar. Demonstrates application of NLP concepts with reviews of key scientific papers.
University Research Group Lab Website
Research group website for the AI and Swarm Intelligence Lab at Göttingen University, supporting work in Swarm Intelligence, Wireless Sensor Networks, and Artificial Intelligence.
ZombieSim
Parallelized simulation exploring HPC concepts. Combines a configurable program with a game engine for visualization, demonstrating practical parallelization techniques.