Skip to content
Layers Deep

AI Engineering Foundations

Go from loops to a retrieval pipeline that answers with evidence, then make it fast and affordable with caching.

What you'll be able to do

  • Trace and debug loop-based code with tests
  • Build intuition for every stage of a RAG pipeline and tune it with evaluation
  • Make caching decisions for AI and web features with explicit tradeoffs

For: Beginners with some code and developers moving into AI features.

At a glance

Modules
7
Ready now
3 lessons
Coming soon
16 lessons
Total time
3 h 10 min

Projects: Document search app, Evaluated RAG assistant

New to programming

Start with First steps when it launches. Until then, Loops includes a short refresher on variables and arrays.

Already write code professionally

Answer the loop questions in the placement check correctly to mark Loops as known and start at RAG.

Take the placement check

Roadmap

  1. 01Coding foundations

    First steps: values and decisions

    Store values, name them well, and let your program choose between paths.

    Optional for beginners
    1. Values, variables, and typesComing soon8 min
    2. Making decisions with conditionalsComing soon8 min

    Milestone: Tip calculator with edge cases. Passing tests for normal and edge-case inputs.

  2. 02Coding foundations

    Iteration and functions

    Repeat work with loops, package it in functions, and organize data in collections.

    1. Loops, step by stepAvailable10 min
    2. Functions that transform dataComing soon9 min
    3. Arrays, objects, and mapsComing soon10 min

    Milestone: Debug a failing loop. All tests pass after at least one failing run, without revealing the solution.

  3. 03AI foundations

    Representations and responsibility

    How embeddings turn meaning into coordinates, and how to use AI responsibly.

    1. Embeddings: meaning as coordinatesComing soon10 min
    2. Responsible use: bias, privacy, and limitsComing soon9 min

    Milestone: Responsible use review. A review covering data, failure modes, and user impact.

  4. 04LLM engineering

    How LLMs read and write

    Tokens, context windows, and a conceptual model of attention.

    1. Tokens and tokenizationComing soon9 min
    2. Context windowsComing soon9 min
    3. Attention, conceptuallyComing soon11 min

    Milestone: Context budget plan. A budget that fits within a stated limit with reasons for each cut.

  5. 05RAG engineering

    Retrieval-augmented generation

    Follow documents from ingestion to grounded, cited answers.

    1. RAG, end to endLocked12 minFinish Loops, step by step first
    2. Ingestion and parsingComing soon10 min
    3. Chunking strategiesComing soon10 min

    Milestone: Evaluate a retrieval pipeline. A passing evaluation run within the context budget (earns Retrieval Evaluator).

  6. 06System design

    Performance and scale

    Caching, queues, and load balancing: the tools for serving more with less.

    1. Caching: hits, misses, and freshnessAvailable11 min
    2. Queues and backpressureComing soon10 min
    3. Load balancingComing soon10 min

    Milestone: Explain a caching tradeoff. A structured tradeoff explanation with no risky decisions (earns Tradeoff Thinker).

  7. 07RAG engineering

    Grounded answers and evaluation

    Generate answers that cite evidence, evaluate them, and analyze failures.

    1. Grounded generation and citationsComing soon11 min
    2. Evaluating RAGComing soon12 min
    3. Failure analysisComing soon11 min

    Milestone: Failure analysis report. A report linking each failure to a pipeline stage.

What the states mean

Completed:
Finished, or marked known by the placement check
Recommended:
The best next lesson, based on prerequisites
Available:
Ready to start
Locked:
Needs a prerequisite first (you can still open it)
Coming soon:
Planned but not built yet