CloudspaceOPTIONAL DEVELOPER EXPLAINERS
Built for the workshop.Use ← → to move between slides. Choose a slide from the learning path above.

Download exercise pack ↓
Day 01 / Extension
Optional

Optional developer deep dives

Extra practice after the core workshop.

These technical animations and cases are extensions. They are not prerequisites for the beginner projects. Return to the programme when you want the core course.
Programme overview →
Day 01 / Concept & relevance
Optional extension

What is source curation?

Understand the concept and its role in your development work.

What it is

Source curation is selecting, checking, and organising the material used to answer a question. A source is the document or excerpt; its authority and version determine how much weight it should carry.

Why it matters

An inherited project often contains old guides, current contracts, and proposed features together. Without curation, a plausible answer can describe the wrong version of the system.

How it works: Identify the question → inspect authority and version → label the source → select relevant evidence.
Learning target: Given the eight source files, label each source’s status and identify the current API contract without treating the proposal as implemented.
Choose evidence before asking AI
Day 01 / Concept & relevance
Optional extension

What is a source-backed system explanation?

Understand the concept and its role in your development work.

What it is

It is an account of how a system works in which each technical claim can be traced to evidence. A citation points to evidence; verification checks that the evidence supports the exact claim.

Why it matters

During onboarding or a code review, an incorrect assumption about permissions or data flow can lead to a wrong implementation. Traceable explanations let another developer check your reasoning.

How it works: Ask a bounded question → locate evidence → connect the workflow → verify every claim.
Learning target: Using current sources, explain one request flow and produce five technical answers whose supporting passages a partner can verify.
A citation is a pointer, not proof
Day 01 / Concept & relevance
Optional extension

What is evidence-led incident triage?

Understand the concept and its role in your development work.

What it is

Triage is the first structured investigation of a problem: record observations, compare them with expected behaviour, and choose the next check. A hypothesis is a possible explanation; it is not yet a verified cause.

Why it matters

A generic “submission failed” message hides useful detail. Developers can narrow the problem faster by examining the request and response before changing code.

How it works: Observe the request → compare the contract → rank causes → define a confirming check.
Learning target: Use the incident evidence to distinguish facts from hypotheses and propose a confirming request with an expected outcome.
From an error to a testable hypothesis
Day 02 / Concept & relevance
Optional extension

What does it mean to ground an answer?

Understand the concept and its role in your development work.

What it is

Grounding connects a response to supplied evidence. Here, the approved reference guide is included in the model’s context. A source ID identifies a reference; it does not prove the response used that reference correctly.

Why it matters

A model’s general knowledge does not establish your project’s current rules. Developers need to constrain project-specific answers and identify claims that exceed the evidence.

How it works: Supply current context → answer within its scope → inspect references → handle missing evidence.
Learning target: Use approved context to answer supported questions, ask useful clarifications, and identify unsupported answers with valid source references.
Answer, clarify, or acknowledge a gap
Day 02 / Concept & relevance
Optional extension

What is a structured response?

Understand the concept and its role in your development work.

What it is

A structured response follows a defined shape, such as a JSON object. A schema specifies fields and types. Validation checks whether the result satisfies the schema and the application’s rules.

Why it matters

A UI needs reliable fields for the answer, sources, and next step. Free-form prose is harder to process, but well-formed JSON can still contain an incorrect answer.

How it works: Define fields → request structured output → validate shape → verify meaning.
Learning target: Apply the provided schema and assess five responses for field types, valid source IDs, factual support, and appropriate clarification.
Two checks for one JSON response
Day 02 / Concept & relevance
Optional extension

What is API integration?

Understand the concept and its role in your development work.

What it is

API integration connects components through requests and responses. The browser handles interaction; the server controls credentials and assembles the Gemini request. Responses may represent success or a service failure.

Why it matters

An assistant can fail even when its prompt is good. Developers must protect credentials, validate input and output, and prevent service errors from becoming misleading answers.

How it works: Input → server request → model response → validation → UI state.
Learning target: Trace one request, locate server-side secret access, and verify empty-input and simulated-failure behaviour without exposing a key.
Follow a request without exposing the key