Senior Programmer

I design original deep learning architectures for biological, clinical, genomic, and other complex data, with research spanning transformers, generative autoencoders, and neural architecture search.

In my independent research, I built HSGO-Transformer to fuse TCGA genomic and clinical data for cancer subtype classification, achieving an AUC-ROC of 0.981 ± 0.007. I design ablation studies and uncertainty-aware validation from the outset rather than relying on convenient aggregate results.

I also developed AEVAE and NEAE, adaptive generative autoencoder architectures for anomaly detection in time-series and network-traffic data, published in IEEE Transactions on Neural Networks and Learning Systems and The Journal of Supercomputing. My publication record includes 10 peer-reviewed papers, including two IEEE Transactions papers.

At Iraq's Communications and Media Commission, I translate applied AI research into deployed production systems. As sole architect and lead engineer for more than 86 active systems, I lead GPU infrastructure scheduling and supervise a mixed research and engineering team.