17 stations from your first API call to running AI in production. Each has short lessons and a check. Open any station to see them.
Role playbooks in six families. Each card shows how long its route is and the skills interviewers ask about most.
Most picked by people starting out
AI Engineer
AI and ML
7 skills on the route
Document ingestion, Embeddings + vector index, Retrieval + reranking
Data Scientist
Data
6 skills on the route
Problem framing and metrics, SQL, Python and pandas, Feature engineering
Data Engineer
Data
6 skills on the route
Source systems and contracts, Ingestion (batch and streaming), Object storage, columnar files and warehouse
ML / MLOps Engineer
AI and ML
6 skills on the route
Feature pipelines and data versioning, Experiment tracking (MLflow, Weights & Biases), Model registry and packaging
Data Analyst
Data
6 skills on the route
Question framing and metric definitions, SQL, Spreadsheets and data cleaning
Backend Engineer
Software
6 skills on the route
API design (REST and OpenAPI), Service language and framework, Relational database (PostgreSQL)
Pick a topic. Every card opens the real lesson, drill or project.
LessonBeginnerLesson of the dayRAG: chunk, retrieve, rerankGive a model the right documents at question time.Start
LessonIntermediateAgents: tools, MCP, memoryModels that call tools in a loop to finish a task.
LessonBeginnerTransformers: vectors, attention, FFNNThe architecture behind every modern LLM.
LessonIntermediateAgentic AI, start to endAll 83 patterns: when to use each, when not to, and how it fails.
DrillIntermediateThe answer is in the docs, but RAG misses itFind why retrieval skips the passage, step by step.
DrillIntermediateAn agent stuck in a tool loopSpot the missing stop condition and fix it.
DrillBeginnerThe LLM bill doubled overnightTrace the cost spike to its cause.
ProjectBeginnerDocs Q&A with citationsBuild RAG that shows its sources.
ProjectIntermediateEval harness with a CI gateGolden sets, LLM-as-judge and a release gate.
ProjectIntermediateTool-calling agent with guardrailsTools, stop conditions and human approval.
ProjectAdvancedDatabricks document pipelineDelta Lake, PySpark and vector search.
PracticeBeginnerPrompt playgroundRun two prompts on a test set and see which scores better.
PracticeIntermediateCoding exercisesChunking, similarity and SQL, checked in the browser.
DemoBeginnerGuided demosStep through RAG, embeddings and agent loops.
Mock interviewIntermediateMock interviewAnswer, get scored, see what to fix.