All topics
Every topic in the Atlas, grouped by domain, plus the 5 tracks that run through them. 88 topics connected by what each one actually requires first.
Tracks
Data Engineer 84h
The least contested entrance to AI work. Ends with someone who can model, move and serve data a team can trust - and who is one short step from RAG, because chunking and ingestion are ETL problems wearing AI costumes.
01 Python02 SQL03 Data Modelling04 Data Pipelines05 Data Warehouse06 Distributed Processing07 Orchestration08 Data EngineeringAI Engineer 137h
The default line. Ends with someone who can design, build, evaluate and operate an LLM-backed product end to end.
01 Python02 Math for AI03 Machine Learning04 Calculus & Optimisation05 Gradient Descent06 Deep Learning07 Neural Networks08 Attention09 Transformers10 Large Language Models11 Linear Algebra12 Embeddings13 RAG14 Prompt Engineering15 AI Agents16 Git & Collaboration17 AI EngineeringRAG Specialist 197h
The shortest commercially valuable line in the Atlas. Skips most of deep learning and goes hard on retrieval quality.
01 Python02 Math for AI03 Linear Algebra04 Calculus & Optimisation05 Machine Learning06 Gradient Descent07 Deep Learning08 Neural Networks09 Attention10 Transformers11 Large Language Models12 Embeddings13 RAG14 Vector Databases15 Vector Search16 Reranking17 Probability & Statistics18 Supervised Learning19 Model Evaluation20 LLM Evaluation21 Chunking22 RAG Evaluation23 Hybrid Search24 Production RAGAgent Builder 163h
Tool use, memory and planning, with agent evaluation treated as a first-class concern rather than an afterthought.
01 Python02 Math for AI03 Calculus & Optimisation04 Machine Learning05 Gradient Descent06 Deep Learning07 Neural Networks08 Attention09 Transformers10 Large Language Models11 Prompt Engineering12 AI Agents13 Tool Use14 Probability & Statistics15 Supervised Learning16 Model Evaluation17 LLM Evaluation18 Agent Evaluation19 Planning & Reasoning20 Multi-Agent SystemsAI Architect 240h
The longest line. Assumes you will build first, then take responsibility for boundaries, cost, governance and failure modes.
01 Python02 Math for AI03 Git & Collaboration04 Linear Algebra05 Calculus & Optimisation06 Machine Learning07 Gradient Descent08 Deep Learning09 Neural Networks10 Attention11 Transformers12 Large Language Models13 Embeddings14 Prompt Engineering15 RAG16 AI Agents17 AI Engineering18 Probability & Statistics19 Supervised Learning20 Model Evaluation21 LLM Evaluation22 Observability & Evals23 Model Serving24 Inference Optimisation25 Cost & Latency26 SQL27 Data Wrangling28 Guardrails & Safety29 Data Strategy30 AI Architecture