AI-10 · Advanced · ENGLISH · PLANNED
Begin with a baseline, trace a model, measure errors and improve it.
Study time is an initial activity estimate, excluding extra reading, revision and extended laboratory or project work. Videos and question bodies have not been produced by adding this outline.
Complete these courses or demonstrate equivalent knowledge. These are recommended prerequisites, not enforced enrolment locks.
YouTube playlist: ARK Artificial Intelligence | Retrieval, Tool Use and AI Applications
Playlist title reserved in the plan; no YouTube playlist or upload has been created yet.
Sign in to save this learning pathPractical outcome: Build a paper design for a source-citing assistant over original notes.
Explain document parsing using a visual and one worked real-world example. Context: Build a paper design for a source-citing assistant over original notes. End with a prediction or check-for-understanding.
Explain chunking using a visual and one worked real-world example. Context: Build a paper design for a source-citing assistant over original notes. End with a prediction or check-for-understanding.
Explain embeddings using a visual and one worked real-world example. Context: Build a paper design for a source-citing assistant over original notes. End with a prediction or check-for-understanding.
Explain vector similarity using a visual and one worked real-world example. Context: Build a paper design for a source-citing assistant over original notes. End with a prediction or check-for-understanding.
Use the first concepts from this section. Build a paper design for a source-citing assistant over original notes. Produce a labelled diagram, trace or calculation; explain each step and check one changed input.
3–6 questions on the preceding first-half concepts only: recall, visual interpretation, application and one plausible misconception. Explain each answer after submission.
Explain hybrid retrieval using a visual and one worked real-world example. Context: Build a paper design for a source-citing assistant over original notes. End with a prediction or check-for-understanding.
Explain reranking using a visual and one worked real-world example. Context: Build a paper design for a source-citing assistant over original notes. End with a prediction or check-for-understanding.
Explain context assembly using a visual and one worked real-world example. Context: Build a paper design for a source-citing assistant over original notes. End with a prediction or check-for-understanding.
Explain citation grounding using a visual and one worked real-world example. Context: Build a paper design for a source-citing assistant over original notes. End with a prediction or check-for-understanding.
Build a paper design for a source-citing assistant over original notes. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Document parsing, Chunking, Embeddings and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Practical outcome: Specify a bounded assistant that cannot silently spend or publish.
Explain structured outputs using a visual and one worked real-world example. Context: Specify a bounded assistant that cannot silently spend or publish. End with a prediction or check-for-understanding.
Explain schema validation using a visual and one worked real-world example. Context: Specify a bounded assistant that cannot silently spend or publish. End with a prediction or check-for-understanding.
Explain tool contracts using a visual and one worked real-world example. Context: Specify a bounded assistant that cannot silently spend or publish. End with a prediction or check-for-understanding.
Explain permission boundaries using a visual and one worked real-world example. Context: Specify a bounded assistant that cannot silently spend or publish. End with a prediction or check-for-understanding.
Use the first concepts from this section. Specify a bounded assistant that cannot silently spend or publish. Produce a labelled diagram, trace or calculation; explain each step and check one changed input.
3–6 questions on the preceding first-half concepts only: recall, visual interpretation, application and one plausible misconception. Explain each answer after submission.
Explain prompt injection using a visual and one worked real-world example. Context: Specify a bounded assistant that cannot silently spend or publish. End with a prediction or check-for-understanding.
Explain sandboxing concepts using a visual and one worked real-world example. Context: Specify a bounded assistant that cannot silently spend or publish. End with a prediction or check-for-understanding.
Explain human approval points using a visual and one worked real-world example. Context: Specify a bounded assistant that cannot silently spend or publish. End with a prediction or check-for-understanding.
Explain time and usd limits using a visual and one worked real-world example. Context: Specify a bounded assistant that cannot silently spend or publish. End with a prediction or check-for-understanding.
Specify a bounded assistant that cannot silently spend or publish. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Structured outputs, Schema validation, Tool contracts and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Practical outcome: Create a held-out evaluation set for a toy retrieval assistant.
Explain retrieval recall using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Explain answer relevance using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Explain source fidelity using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Explain unsupported claims using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Use the first concepts from this section. Create a held-out evaluation set for a toy retrieval assistant. Produce a labelled diagram, trace or calculation; explain each step and check one changed input.
3–6 questions on the preceding first-half concepts only: recall, visual interpretation, application and one plausible misconception. Explain each answer after submission.
Explain adversarial evaluation using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Explain latency using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Explain caching using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Explain observability using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Explain regression suites using a visual and one worked real-world example. Context: Create a held-out evaluation set for a toy retrieval assistant. End with a prediction or check-for-understanding.
Create a held-out evaluation set for a toy retrieval assistant. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Retrieval recall, Answer relevance, Source fidelity and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Combine all three sections: Create a held-out evaluation set for a toy retrieval assistant. Submit an original worked artifact, evidence or tests, assumptions and a reflection. Rubric: correctness, method, practical usefulness, reproducibility and communication. Self-reviewed, not certified.
6–12 balanced questions sampling every section of Retrieval, Tool Use and AI Applications, including interpretation, calculation or reasoning, a misconception and a new application. Passing is 70%; review errors before retrying.
Data experiments, evaluations, model tracing and reproducible projects. Guided work leads to an independent attempt, with a diagram, explanation, trace, calculation or tested artifact. Question bodies are generated when opened in your private learning path. Programming submissions are stored as text; the platform does not execute submitted code.
Use original or appropriately licensed datasets. Practical science activities use models or supplied data unless a qualified supervisor has reviewed a physical experiment.
Open the interactive practice sampler →