Learning paths
Pick a destination, not a playlist
Each path sequences the articles, projects, tools and courses already on this site into a route toward a specific role.
AI Engineer
From prototype to a language-model system that survives production
- Measure retrieval and generation separately and act on the difference
- Build a golden set and a regression gate
- Design agents that fail safely and within budget
ADVANCEDDifficulty: ADVANCED7 steps~12 weeks
Business Analyst
Turn ambiguity into decisions, with AI doing the mechanical work
- Scope a vague request into a testable requirement
- Run and verify an analysis end to end
- Write a recommendation someone can act on
BEGINNERDifficulty: BEGINNER4 steps~8 weeks
Data Analyst
SQL, analysis and a portfolio that gets interviews
- Write SQL that goes beyond SELECT with confidence
- Analyse honestly and communicate a recommendation
- Build three portfolio projects worth showing
BEGINNERDifficulty: BEGINNER5 steps~10 weeks
Data Engineer
Pipelines, modelling and orchestration — starting on your laptop
- Build idempotent, incremental, tested pipelines
- Model data for consumption rather than storage
- Orchestrate with real dependencies and schedules
BEGINNERDifficulty: BEGINNER5 steps~14 weeks
Power BI / Fabric Professional
From report developer to platform engineer
- Write SQL and reason about query plans
- Design a medallion architecture with rules per layer
- Diagnose Direct Lake fallback and model performance
INTERMEDIATEDifficulty: INTERMEDIATE6 steps~14 weeks
Automation Builder
Automations that still work in six months
- Choose what to automate on evidence
- Build idempotent workflows with real error handling
- Design a review queue people will actually use
INTERMEDIATEDifficulty: INTERMEDIATE4 steps~8 weeks
AlgoStrategy Researcher
Understand it. Code it. Backtest it. Honestly.
- Recognise the five biases that inflate backtests
- Build point-in-time data handling
- Model costs that change which strategies work
ADVANCEDDifficulty: ADVANCED6 steps~12 weeks