Resume
Mousa Shaya
Summary
Senior software architect and ML researcher with 45+ years of depth across real-time embedded systems, enterprise Java/cloud platforms, and modern deep learning pipelines. Seven patents in deployed biosignal pattern recognition (HP/Agilent, ICU deployment worldwide). Current work spans SAR remote sensing, biosignal ML, biodiversity informatics, and security operations tooling. Founder, Shaya Software and Systems LLC.
Skills
ML & Data Science: Python, PyTorch, scikit-learn, XGBoost, SHAP, DINOv2, BioCLIP, ViT, UMAP, HDBSCAN, SAR/remote sensing (Sentinel-1 SLC/GRD), PSG/EEG/ECG signal processing, HuBERT, Wav2Vec 2.0, MAFFT, IQ-TREE2, phylogenetic inference
Security & Infrastructure: OpenSearch/Elasticsearch, Wazuh SIEM, MITRE ATT&CK, NIST 800-171, CMMC, Python/Starlette, REST APIs
Enterprise Java & Cloud: Java, Spring Boot, AWS (DynamoDB, RDS, ElastiCache), JPA, Hibernate, MyBatis, Tomcat, JMS, SOAP/REST, JSON, JDBC
Databases: PostgreSQL, MySQL, SQL Server, DynamoDB, HSQL, SQLAnywhere
Methods & Tools: Git, Maven, Scrum, Kanban, UML, Root Cause Analysis, API governance, microservices architecture, database design (Erwin)
Experience
Founder & Principal Consultant
ML Research
-
SAR Automatic Target Recognition — systematic generalization study across depression angles, sensor configs, and datasets (ATRNet-STAR, MSTAR). Six contributions including +30 pp accuracy improvement from multi-angle training, 98%+ few-shot accuracy with 5 reference images per class. Under review, IEEE Geoscience and Remote Sensing Letters.
-
SAR Ship Classification — ship type classification from Sentinel-1 SLC imagery with a multi-signal AIS label resolver (AIS type, ship length, GFW classification, name keywords) resolving 87% of 2,088 vessels without manual annotation. DINOv2 ViT-S/14 classifier on resolved labels. Under review, IEEE.
-
PhysioNet 2026 Challenge — open-source algorithms for cognitive impairment screening from polysomnography (PSG) — EEG, ECG, and other physiological signals. Training data from five US academic medical centers.
-
Lichen Image Classification — BioCLIP ViT-L/14 + ArcFace metric learning classifier for species-level lichen identification. Two-stage training on iNaturalist and expert-labeled Lichen Consortium data (~15,000 images, ~10,000 species). ~95% top-1 accuracy. Collaboration with Dr. Frank Bungartz, Arizona State University.
Security Operations Tooling
- Designed and developed a browser-based SOC dashboard for Wazuh/OpenSearch SIEM environments: MITRE ATT&CK kill chain timeline visualization, NIST 800-171 and CMMC compliance gap detection, SCA policy assessment, vulnerability management, geo-IP enrichment, multi-source log normalization (Wazuh 4/5, Elastic Security ECS), multi-tenant architecture, and scheduled PDF reporting. Python/Starlette backend, OpenSearch, single-file HTML/CSS/JS frontend.
Biodiversity Database Engineering (ASU Lichenology Consortium, Nov 2024–Present)
-
Cleaned the Consortium database: removed 35,000 non-lichen taxa, added 60,000 validated taxa from Mycobank and Index Fungorum, updated 200,000 records with external source references. Implemented cross-source correlator (LIAS, Mycobank, Index Fungorum) with lichen filter and taxonomic hierarchy validation. Contributor to Symbiota open-source platform.
-
Phylogenetic analysis: parsed ~51K GenBank DNA records (8 loci), MAFFT alignment, trimAl trimming, IQ-TREE2 maximum likelihood inference, LIAS trait mapping onto trees.
-
UMAP + HDBSCAN clustering over LIAS DELTA morphological data (10,709 species, 880 chemical compounds); interactive three-panel cluster browser.
Independent Consultant
-
COOP FCL — UKG/JDE Integration (2024–2025): Integrated UKG Activities module with Oracle JD Edwards Orchestrator APIs for employee activity tracking in a refinery environment. Built all UKG Activity structures and bidirectional sync; Java application deployed on-premise.
-
COOP FCL — Automated Testing (2025): Developed UKG API test harness using UAT environment to punch in/out employees across scenarios (daily OT, weekly OT, stat pay, lunch deduct, RBS, stat holiday). Automated hourly comparison to expected hours; production config verification.
-
BJs — Workforce Management (2023): Application consulting: Wi-Fi automation, clock configuration, dataviews, attestation, TestAssured, business structure, employee groups. UAT refresh and UAT-to-PRD configuration transfer via SDM.
Principal Architect
-
One of two architect leads directing 11 teams (~110 engineers, testers, team leads) across Belarus and Lowell to migrate on-premise Activity product to cloud-based Dimensions platform.
-
Lead architect for the Kiosk product: secure, limited Dimensions access via badge ID, full authentication, and federated login. Led three teams across Noida and Lowell.
-
Lead architect for production SQL performance remediation across 1,000+ customer clusters; reduced database CPU utilization significantly through query plan analysis and index redesign.
-
Responsibilities across all projects: stakeholder and product owner requirements, high-level architecture and database design, API governance, UX standards, security reviews, production incident response, microservices migration.
Team Lead & Scrum Master
Team lead and Scrum Master for Control Point releases. Backlog grooming, SQL Server normalization, and data flow optimization.
Principal Software Development
-
Scaled legacy SQL Server solution to several hundred million documents using file groups, partitioned tables, index optimization, and execution plan analysis.
-
Designed and prototyped a federated PostgreSQL + HSQL distributed document analysis system for scanning billions of documents; Java Spring Boot, Tika, OpenNLP, REST architecture.
Software Architect, Hotel Search
Extended hotel search with 12+ new providers (SOAP/JSON/XML); booking/cancellation for two new providers. Aggregated Hadoop/Hive performance metrics into MySQL for monitoring. Created public REST APIs; implemented AES encryption of sensitive data; optimized inventory priming for faster user searches.
Principal Software Development Engineer / Project Lead
-
Architected and delivered payment policy service for Nokia mobile commerce: credit card and 150 operator billing methods worldwide. Delivered in 4.5 months with limited resources. AWS, Spring, REST, JSON, JPA, Amazon RDS.
-
Designed NextGen in-app purchasing for all Nokia devices: simplified payment API, improved UX and throughput on high-latency networks, Spring/JBoss, MySQL.
-
Product restoration service for purchased content across devices; DynamoDB with global indices and ElastiCache for eventual consistency.
Education
Northeastern University — Boston, MA
BSEE / MSEE — Communications Systems and Digital Signal Processing
Patents (HP / Agilent Technologies)
- US 4,664,116 — Pace Pulse Identification Apparatus
- US 4,832,041 — Pace Pulse Eliminator
- US 4,539,999 — Method and Device for Subtracting a Pacer Signal from an ECG Signal
- US 4,838,278 — Paced QRS Complex Classifier
- US 5,447,164 — Interactive Medical Information Display System and Method for Displaying User-Definable Patient Events
- A61B5/7264 — Classification of Physiological Signals Using Neural Networks, Statistical Classifiers, and Expert Systems
Affiliations
- ASU Research Affiliate — Arizona State University
- Member — AFCEA Lexington-Concord Chapter
- SAM.gov Registered Vendor — UEI YK29WQZ22EM3
- Registered Expert — GLG Expert Network / Catalant
- Member — American Bryological and Lichenological Society (ABLS)
- Contributor — Consortium of North American Lichen Herbaria (CNALH) / Symbiota