## Principal Consultant — Machine Learning & Signal Processing Shaya Software and Systems LLC is an independent ML consultancy founded on rare depth in real-time signal processing, physiological waveform analysis, and applied machine learning. Before deep learning frameworks existed, I designed ECG pattern recognition algorithms deployed in ICUs worldwide — earning 7 patents at HP/Agilent. That foundational rigor in signal decomposition, feature engineering, and classification under noise now directly informs modern ML work in medical biosignal processing and biodiversity informatics. Current research applies transformer-based vision models, metric learning, and phylogenetic analysis to lichen biodiversity in collaboration with Arizona State University — toward academic publication in methods-focused venues. ## Available For - **Medical ML & biosignal processing** — PCG/ECG/PPG classification, physiological waveform analysis, clinical-grade validation - **Biodiversity & ecological ML** — species image classification, phylogenetic pipelines, biodiversity database engineering - **ML systems architecture** — data pipelines, model deployment, production-grade design ## Current Work Active research with Dr. Frank Bungartz (ASU) on automated lichen classification: BioCLIP ViT-L/14 + ArcFace metric learning pipeline, UMAP/HDBSCAN clustering across 10,000+ species, multi-locus phylogenetic inference from GenBank DNA data, and interactive cluster visualization tooling. Parallel investigation into PCG (heart sound) classification using mel-spectrogram + ViT architectures and pretrained audio foundation models. Development workflow incorporates Claude Code (Anthropic) for accelerated ML pipeline implementation, consistent with current industry practice for senior ML engineers. [View Research →](/research/){: .btn .btn--primary} [Consulting →](/consulting/){: .btn .btn--inverse}

Recent Posts