Research with evidence
Summarize scholarly and technical literature, label uncertainty, distinguish evidence from speculation and maintain source provenance. Research accuracy must be evaluated.
AEVYRIUM / Intelligent systems
AI systems research & engineering
Knowledge. Tools. Reasoning.
The Academy defines broad shared foundations and specialist study domains. The skills below are target competencies and development tracks; individual competence, model quality and real-world tool access require separate evidence.
NINE SHARED FOUNDATIONS
Core AI capability architecture.
Engineering and software
Work toward programming, code analysis, API design, test generation, reproducible builds, observability and debugging with approved repositories and tool permissions.
Multidisciplinary reasoning
Apply mathematics and scientific reasoning to physics, engineering, biological research, economic models and complex project analysis.
Multi-agent collaboration
Coordinate specialized digital roles around a shared objective while preserving identity boundaries, task accountability and independent verification.
Retrieval and memory boundaries
Develop versioned knowledge retrieval, project-context awareness, records of evidence, identity persistence and privacy-aware separation of memory.
Communication and multimodal interfaces
Design natural-language, speech, image, video and 3D workflows, with each modality requiring separately verified model and user-interface behavior.
Planning and operations
Support roadmaps, resource allocation, procurement research, production scheduling and structured decision support without unauthorized autonomous purchasing.
Owner-authorized tools
Introduce integrations through authenticated permission scopes, allowlists, limits and human approval for consequential actions.
Evaluation and continuous improvement
Use regression suites, experiments, model/dataset provenance, training manifests and controlled deployment gates; never equate configured curricula with trained competence.
THIRTY KNOWLEDGE DOMAINS
Specializations spanning the sciences and enterprise.
The source catalog assigns these 30 named knowledge domains to the avatars as reference and practice areas. Training assignments are not competency certifications.
AI and computer systems
- Advanced programming
- AI and model engineering
- Multi-agent coordination
- Operating systems, databases and networks
- Defensive cybersecurity
Engineering and energy
- Robotics and embodied control
- Physics
- Chemistry and materials
- Electrical engineering
- Energy research and storage
Earth and space
- Aerospace engineering
- Astronomy and planetary science
- Environment and restoration
- Civil engineering and infrastructure
- Geospatial intelligence and satellite telemetry
Life and human systems
- Genomics and bioinformatics
- Medical science literacy
- Psychology and communication
- Education and assessment
- Humanitarian and social-impact systems
Business and society
- Finance and economic modeling
- Business and operations management
- Legal research and governance
- History, archives and provenance
- Languages, speaking and translation
Foundations and creative work
- Advanced mathematics
- Scientific reasoning and reproducibility
- Multimedia and creative production
- Simulation and digital twins
- Long-horizon feasibility research
LEARNING PHILOSOPHY
Curriculum is not the same as mastery.
Evidence standard: Lessons, tutorials and research data must pass provenance and licensing review. Competence should be assessed through held-out problems, safety tests, versioned evaluation results and independent domain expert review.