Assessed edition · 1 Oct 2026
A field guide to AI capabilities / 2026
Know the landscape.
Choose your next move.
Explore the concepts, understand the distinctions, and find where learning should stop. An evidence-led reading companion to the Skills Atlas.
Start with the RAG brief47 of 47 concepts
Models & learning
10- Direct Preference Optimization2 taxonomy axesOutside this profile
- In-Context Learning1 taxonomy axisExtend
- Large Language Models1 taxonomy axisCover the essentials
- Long-Context Modeling1 taxonomy axisRun a learning trial
- LoRA / QLoRA1 taxonomy axisOutside this profile
- Reasoning Models1 taxonomy axisRun a learning trial
- Structured LLM Outputs1 taxonomy axisCover the essentials
- Supervised Fine-Tuning (SFT)1 taxonomy axisOutside this profile
- Test-Time Compute Scaling1 taxonomy axisRun a learning trial
- Training Data Curation1 taxonomy axisOutside this profile
Knowledge & retrieval
9- Agentic RAG1 taxonomy axisDefer with a trigger
- Dense Retrieval1 taxonomy axisCover the essentials
- Embedding Models1 taxonomy axisRun a learning trial
- Hybrid Search2 taxonomy axesRun a learning trial
- Multimodal RAG2 taxonomy axesDefer with a trigger
- Retrieval-Augmented Generation1 taxonomy axisExtend
- Search Re-Ranking1 taxonomy axisRun a learning trial
- Vector Databases2 taxonomy axesCover the essentials
- Vector Indexing2 taxonomy axesDefer with a trigger
Agents & tools
8- Agent Memory Systems1 taxonomy axisDefer with a trigger
- Agentic Planning & Task Decomposition2 taxonomy axesDefer with a trigger
- AI Agent Design1 taxonomy axisCover the essentials
- AI-Assisted Development1 taxonomy axisRun a learning trial
- Code Execution Agents1 taxonomy axisDefer with a trigger
- FastAPI1 taxonomy axisRun a learning trial
- Model Context Protocol1 taxonomy axisDefer with a trigger
- Multi-Agent Coordination Patterns1 taxonomy axisDefer with a trigger
Quality & evaluation
6Safety & security
5Platform & performance
9- Amazon Bedrock1 taxonomy axisDefer with a trigger
- Docker1 taxonomy axisCover the essentials
- Edge AI1 taxonomy axisOutside this profile
- GPU Acceleration1 taxonomy axisOutside this profile
- Inference Optimization1 taxonomy axisExtend
- LLM Inference Serving2 taxonomy axesOutside this profile
- Model Quantization1 taxonomy axisOutside this profile
- Speculative Decoding2 taxonomy axesOutside this profile
- vLLM1 taxonomy axisOutside this profile
The six areas are provisional navigation groups, not a complete or mutually exclusive model of AI skills. A concept’s position is not a ranking.