Genomics and longevity research
Sequence analysis, variant interpretation literature, molecular biology education, reproducible computational methods and medical-science research with appropriate expert oversight.
AEVYRIUM / Intelligent systems
AI systems research & engineering
Science supported by computational evidence.
AEVYRIUM AI explores machine learning, engineering simulation and source-grounded scientific research. We clearly separate demonstrated software experiments from future scientific objectives.
COMPUTATIONAL GENOMICS
A documented research prototype.
94.28%Held-out splice-junction classification accuracy
297Held-out test sequences in the source report
0.929Approximate macro-F1 on three classification labels
What was built
A reproducible, CPU-trained multinomial logistic-regression classifier for splice-junction categories, using 60-base windows from an established primate-derived UCI dataset. The model and training evidence are recorded in the software repository.
What the results do not show
These figures do not measure human-genome coverage, clinical interpretation, disease risk, gene-editing outcomes or validated variant effects. Heavy genomic foundation models have not been installed by this experimental baseline.
Research status: Software prototype and test evidence, not a medical device or diagnostic service.
RESEARCH PORTFOLIO
Planned scientific directions.
Robotics and physical intelligence
Simulation-driven perception and motion planning, embedded sensing, controls and potential ORA Robotics applications.
Aerospace and space systems
Research into spacecraft systems, navigation, remote sensing, resource discovery, planetary sciences and mission planning.
Energy and materials
Physics, chemistry, materials engineering, storage and energy-harvesting concepts subject to feasibility tests.
Climate and Earth restoration
Data-informed environmental monitoring, resilient infrastructure, clean water systems, restoration planning and responsible resource use.
Human systems and education
Accessible learning systems, human-impact research, digital preservation, knowledge archives and social-impact operations.
RESEARCH INTEGRITY
Scientific claims must be reproducible.
Research outputs require sourced claims, dataset permissions, documented evaluation protocols, explicit uncertainty and independent scrutiny where consequential. High-risk life-science and infrastructure applications require qualified professionals and regulatory review.