Guide to AI for Modern Software Engineering

Beschreibung

Machines now write a great deal of software. Teams are not shipping better systems, and in some measured cases they are shipping more slowly. This textbook/guide explains why, and what to do about it.

Artificial intelligence rarely removes engineering work; it relocates it. When generating code becomes cheap, the constraint moves to reviewing it. When detecting anomalies becomes cheap, the constraint moves to deciding which ones matter. This book follows that pattern across four disciplines usually taught apart — DevOps, DevSecOps, MLOps and AIOps — and treats them as one subject, because engineers work across all four over a single system. Twenty chapters cover intelligent delivery pipelines, security testing and compliance, the model lifecycle, observability and incident response, and predictive capacity and cost.

Topics and features:

•Grounded throughout in one running example at realistic scale, with worked arithmetic the reader can follow and challenge

•Full instructor materials: teaching plans, case studies, assessments with rubrics, slide decks with speaker notes, and figure alt text for accessibility

•More than 200 original figures, each explained element by element rather than left to the reader

•A public companion repository of runnable code and end-to-end deliverables

•Six appendices: tool comparison, Python examples, regulatory quick reference, maturity checklists, glossary and further reading

•Honest about evidence—with each chapter stating its limits, and the closing chapter testing its own forecasts

The book is written for postgraduate and final-year undergraduate students on modules in software engineering, DevOps, secure development and machine learning operations. The content also will appeal to software, security, machine learning and site reliability engineers in practice, as well as engineering leaders deciding what to fund.

Muthu Ramachandran is Principal Research Consultant at Forti5 Tech Ltd, UK, and Visiting Professor Extraordinarius at the University of South Africa. He holds a PhD from Lancaster University, has more than 35 years in software engineering research and practice, and has supervised more than 30 doctoral completions.

Machines now write a great deal of software. Teams are not shipping better systems, and in some measured cases they are shipping more slowly. This textbook/guide explains why, and what to do about it.

Artificial intelligence rarely removes engineering work; it relocates it. When generating code becomes cheap, the constraint moves to reviewing it. When detecting anomalies becomes cheap, the constraint moves to deciding which ones matter. This book follows that pattern across four disciplines usually taught apart — DevOps, DevSecOps, MLOps and AIOps — and treats them as one subject, because engineers work across all four over a single system. Twenty chapters cover intelligent delivery pipelines, security testing and compliance, the model lifecycle, observability and incident response, and predictive capacity and cost.

Topics and features:

•Grounded throughout in one running example at realistic scale, with worked arithmetic the reader can follow and challenge

•Full instructor materials: teaching plans, case studies, assessments with rubrics, slide decks with speaker notes, and figure alt text for accessibility

•More than two hundred original figures, each explained element by element rather than left to the reader

•A public companion repository of runnable code and end-to-end deliverables

•Six appendices: tool comparison, Python examples, regulatory quick reference, maturity checklists, glossary and further reading

•Honest about evidence—with each chapter stating its limits, and the closing chapter testing its own forecasts

The book is written for postgraduate and final-year undergraduate students on modules in software engineering, DevOps, secure development and machine learning operations. The content also will appeal to software, security, machine learning and site reliability engineers in practice, as well as engineering leaders deciding what to fund (who will find the cost and capacity material directly applicable).

Muthu Ramachandran is Principal Research Consultant at Forti5 Technologies Ltd, United Kingdom, and Visiting Professor Extraordinarius at the University of South Africa. He holds a PhD from Lancaster University, has more than thirty-five years in software engineering research and practice, and has supervised more than thirty doctoral completions.

Autor*in

Muthu Ramachandran

Themen in »Guide to AI for Modern Software Engineering«

AI-native software engineering DevOps DevSecOps MLOps AIOps continuous delivery AI-assisted development security testing vulnerability assessment compliance by design model lifecycle observability incident management root-cause analysis capacity planning

Stimmen zu »Guide to AI for Modern Software Engineering«

Details

Link teilen