Quality
Quality is the degree to which an output, system, process, or outcome satisfies relevant criteria or exhibits desirable properties.
Designing AI Systems
A score is a numerical or categorical value representing how well an observed operation or output performed against an evaluation criterion.
- How it is used
- Treats scores as production quality signals alongside operational telemetry and as inputs to experimentation.
- How deep it goes
- Defines score values, sources, and attachment points, then configures asynchronous scoring rules, sampling, and alerts.
- Examples
- Configures a patient intake rule with a helpfulness judge, required-element checks, retrieval relevance, and alert thresholds.
- What you can do after
- Operate/configure production scoring rules with sampling, evaluators, and alert thresholds.
AI Recommender Systems
The recommender-system stage that assigns predicted relevance or affinity scores to eligible candidate items.
- How it is used
- Assigns relevance signals to filtered candidates before ordering combines scores with business logic.
- How deep it goes
- Works through adding normalized popularity to similarity-scored movie candidates, while leaving further signals as extensions.
- Examples
- Uses movie similarity and rating-count popularity, plus predicted purchase probability for co-purchased products.
- What you can do after
- Build it from scratch by adding popularity scores to filtered candidates and passing the enriched candidates to ordering.
Build AI Drug Discovery Pipelines
A scoring function is a function that assigns a numerical value to a candidate according to specified criteria.
- How it is used
- Ranks compounds, molecular poses, or generated structures to prioritize promising candidates for screening, docking, or experimental assessment.
- How deep it goes
- Explains scoring intuitively and operationally, including ranking, enrichment evaluation, docking terms, diagnostics, and validation limits.
- Examples
- Uses EGFR pIC50 ranking, Hinge Binders Library prioritization, enrichment factors, and Vina docking affinities for concrete examples.
- What you can do after
- Evaluate candidates by ranking scores, calculate enrichment factors, inspect docking diagnostics, and recognize when scores require further validation.
Read how Build AI Drug Discovery Pipelines explains quality →
Build High Value Software Systems
The degree to which software satisfies its requirements and is free from defects.
- How it is used
- Uses quality to locate where defects enter the delivery loop and how to catch them earlier.
- How deep it goes
- Develops quality through phase-by-phase diagnostic questions and early-detection practices, without defining a numerical quality metric.
- Examples
- Illustrates late defects with an API endpoint and test-suite problems with slow, flaky tests and manual reruns.
- What you can do after
- Diagnose/debug with it by tracing defects to their origin, then choosing checks that catch them earlier.
Read how Build High Value Software Systems explains quality →
Think Like a Software Engineering Manager
Code quality is the degree to which source code satisfies standards for correctness, maintainability, clarity, and related engineering attributes.
- How it is used
- Guides engineering managers in setting coding expectations, reviewing work, coaching performance, and reducing technical debt.
- How deep it goes
- Explains standards, reviews, modularity, measurement, coaching, and technical-debt questions without detailing automated-tool implementation.
- Examples
- Uses naming and formatting standards, peer-review comments, revision counts, modular components, coaching records, and technical-design questions.
- What you can do after
- Apply code-quality practices by setting standards, reviewing changes, tracking feedback, coaching gaps, and prioritizing debt reduction.
Read how Think Like a Software Engineering Manager explains quality →
Build a Large Language Model (From Scratch)
A numeric evaluation result used here to summarize how well the fine-tuned model's responses are judged by another model.
- How it is used
- Quantifies fine-tuned model performance on the test set.
- How deep it goes
- Defines a 0–100 integer rating and shows how to compute an average score, but stops short of deeper evaluation theory.
- Examples
- Uses test-set examples, per-entry prompts, and an averaged score such as 50.32.
- What you can do after
- Apply it to score responses and compute a benchmark average for comparison.
Read how Build a Large Language Model (From Scratch) explains quality →
Lead Developer Career Guide
The degree to which software satisfies requirements and achieves desirable properties such as reliability, maintainability, and correctness.
- How it is used
- Prioritizes quality while planning and delivering software projects.
- How deep it goes
- Connects quality to standards, reviews, testing, maintainability, and workload, without describing a scoring method.
- Examples
- Uses a reusable CSV library, a React starter project, and workplace examples of deadlines, resource constraints, and burnout.
- What you can do after
- Practice quality-focused delivery by reviewing code, testing thoroughly, and planning realistic workloads and resources.
Architecture Modernization: Socio-technical alignment of software, strategy, and structure
Code health is the maintainability and internal quality of a codebase, including properties that support safe and efficient change.
- How it is used
- Protects the ability of strategic and high-value domains to evolve quickly over the long term.
- How deep it goes
- Explains business and organizational importance through architecture, domain boundaries, dependencies, and team incentives, but offers no refactoring procedure.
- Examples
- Uses strategic IT, a decisive core, advanced architecture, rich domain models, and legacy code as contrasting contexts.
- What you can do after
- Evaluate code health as a business enabler whose neglect slows change and increases long-term cost and risk.
Vibe Engineering
Quality is the degree to which a system's outputs satisfy required correctness, usefulness, and user expectations.
- How it is used
- Positions output quality within workflow choices that balance latency, cost, task complexity, and user expectations.
- How deep it goes
- Compares workflow and quantization tradeoffs, but stops short of defining a universal quality metric or evaluation protocol.
- Examples
- Uses customer complaints, model-workflow assignments, quantization levels, acceptance rates, completions, functions, and generated test suites.
- What you can do after
- Evaluate model configurations against representative workflows and choose where to spend latency, cost, or quality.
Software Testing with Generative AI
The degree to which a language-model response is legible, useful, accurate, and appropriate for its intended purpose.
- How it is used
- Evaluates language models against the requirements of a specific problem rather than relying on general performance indicators.
- How deep it goes
- Defines practical criteria and recommends comparative experimentation, but stops short of a formal scoring framework.
- Examples
- Uses comparative code-generation results from OpenAI GPT models and Google Gemini to illustrate task-specific quality.
- What you can do after
- Evaluate and select models by comparing their responses to prompts for a defined practical problem.
Read how Software Testing with Generative AI explains quality →
Also covered in
- Introduction to Generative AI, Second Edition
- Hidden InfluencesIn early access (MEAP)
- Build Python Web Apps with Streamlit
- "Looks Good to Me"
- AI-Powered Developer: Build software with ChatGPT and Copilot
- Learn AI-Assisted Python Programming, Second Edition
- Quantum Programming in Depth
- Coding with AI
- AI-Powered Search
- Build and Orchestrator in Go (From Scratch)
- Terraform in Depth: Infrastructure as Code with Terraform and OpenTofu
- 100 Java Mistakes and How to Avoid Them
What to know first
- language modelA language model is a computational model that represents patterns in linguistic sequences to assign probabilities, predict, generate, or evaluate language.
- promptA prompt is text or other input supplied to an AI model to provide instructions, context, or other information that guides its response or generated output.
- artificial intelligenceArtificial intelligence is the field of computing concerned with creating systems that perform tasks associated with intelligence, such as learning, reasoning, perception, and decision-making.
- accuracyAccuracy is the degree to which a result agrees with a specified truth, reference, or criterion, commonly measured as the proportion of evaluated predictions that are correct.
Related concepts
- trajectoryuses · A trajectory is an ordered record of states, actions, events, or operations showing how an agent or system progresses through an execution or process.
- forecasting qualityincludes · The overall degree to which forecasts meet desired performance criteria, including accuracy and bias.
- code coveragecontrasts with · Code coverage is a measure of which portions or behaviors of a program are exercised during testing.
Questions about quality
What is quality?
Quality is the degree to which an output, system, process, or outcome satisfies relevant criteria or exhibits desirable properties.
Which Manning books explain quality?
Designing AI Systems, AI Recommender Systems, Build AI Drug Discovery Pipelines, Build High Value Software Systems, Think Like a Software Engineering Manager, Build a Large Language Model (From Scratch), Lead Developer Career Guide, Architecture Modernization: Socio-technical alignment of software, strategy, and structure and 2 more.
What should I know before learning quality?
Manning books that teach quality usually introduce language model, prompt, artificial intelligence and accuracy first.
What is the difference between quality and code coverage?
Quality is the degree to which an output, system, process, or outcome satisfies relevant criteria or exhibits desirable properties. Code coverage is a measure of which portions or behaviors of a program are exercised during testing.
What kinds of quality are there?
Kinds of quality include api design quality, forecast quality, forecasting quality and image quality.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.









