We may earn an affiliate commission when you visit our partners.
Course image
Pearson
Enroll now

What's inside

Syllabus

Save this course

Create your own learning path. Save this course to your list so you can find it easily later.
Save

Career center

Learners who complete Securing Generative AI will develop knowledge and skills that may be useful to these careers:
AI Security Engineer
An AI Security Engineer specializes in designing, implementing, and maintaining robust security measures for artificial intelligence systems, including large language models and Retrieval Augmented Generation. This role is crucial for preventing vulnerabilities and ensuring the integrity and confidentiality of AI applications. The Securing Generative AI course provides a comprehensive exploration of critical security measures, directly aligning with the core responsibilities of an AI Security Engineer. Learners gain practical knowledge of current AI threats such as prompt injection and insecure output handling, alongside mitigation strategies and "secure by design" principles. This course helps build a strong foundation in safeguarding AI technologies in production environments, crucial for success in this highly specialized and rapidly evolving field. For those aspiring to be an AI Security Engineer, this particular course offers specific insights into securing vector databases and understanding orchestration libraries, directly contributing to practical deployment skills.
Machine Learning Security Researcher
A Machine Learning Security Researcher investigates novel security vulnerabilities and develops innovative countermeasures for machine learning systems and algorithms. This role requires a deep understanding of how AI threats manifest and how to architect resilience. The Securing Generative AI course offers a detailed exploration of generative AI security, covering threats like training data poisoning, model denial of service, and supply chain vulnerabilities. Learners are introduced to critical considerations and mitigations necessary to reduce risk in AI system development processes. This specific course can help a researcher understand the attack surface of contemporary AI, including LLMs and RAG implementations. It also delves into red team AI models, a key activity for a Machine Learning Security Researcher, providing a foundation for identifying and addressing emerging security challenges in artificial intelligence.
Red Team Artificial Intelligence Specialist
A Red Team Artificial Intelligence Specialist is responsible for simulating sophisticated cyberattacks to test the security posture of AI systems, uncovering vulnerabilities before malicious actors can exploit them. This role is inherently focused on proactive threat discovery and validation. The Securing Generative AI course directly addresses "Red Team AI models," providing an invaluable perspective for this specialization. Learners gain an understanding of various AI threats, including prompt injection, insecure output handling, and model theft, which are prime targets for red team exercises. This course helps individuals develop the practical knowledge needed to identify weaknesses in generative AI implementations, including those leveraging RAG, orchestration libraries like LangChain, and vector databases. For an aspiring Red Team Artificial Intelligence Specialist, this course offers the foundational understanding of adversarial techniques and defensive strategies specific to AI.
Chief Information Security Officer
A Chief Information Security Officer, often requiring an advanced degree, is an executive responsible for an organization's overall information security strategy and implementation. This leadership role involves setting policies, managing risks, and ensuring the protection of all digital assets. As generative AI becomes integrated into critical business operations, understanding its security posture is paramount for a CISO. The Securing Generative AI course offers a comprehensive exploration of crucial security measures necessary for AI deployment and development, addressing critical considerations and mitigations to reduce organizational risk. It emphasizes "secure by design" principles and building organizational structures that prioritize security, which are key strategic objectives for a Chief Information Security Officer. This course helps develop an executive-level understanding of AI threats, LLM security, and how to safeguard AI technologies in production, enabling informed strategic decisions and oversight for enterprise AI adoption.
Cybersecurity Consultant
A Cybersecurity Consultant advises organizations on how to protect their information systems and assets from cyber threats, often developing strategies to mitigate risks. As AI adoption grows, consultants must guide clients in securing their generative AI deployments. The Securing Generative AI course offers a comprehensive exploration into the crucial security measures necessary for deploying AI, directly supporting the advisory function of a consultant. Learners are introduced to "secure by design" principles, focusing on security outcomes and building organizational structures that prioritize security, which are key for effective consulting engagements. This course helps a Cybersecurity Consultant understand AI threats like prompt injection and training data poisoning, and provides practical knowledge of industry frameworks and best practices to safeguard AI technologies in production environments, enabling them to offer expert, tailored advice.
DevSecOps Engineer
A DevSecOps Engineer integrates security practices into the continuous integration and continuous delivery pipelines, ensuring security is embedded from development through operations. As generative AI models become central to software products, securing their lifecycle is critical. The Securing Generative AI course emphasizes "secure by design" principles and offers comprehensive insights into safeguarding AI technologies in production environments. Learners gain practical knowledge about mitigating vulnerabilities like supply chain vulnerabilities and model denial of service, which are direct concerns for DevSecOps professionals managing AI deployments. This course helps a DevSecOps Engineer understand how to implement security controls for large language models and RAG implementations, including securing vector databases and orchestrating libraries. For those in DevSecOps, this particular course provides the necessary expertise to ensure that AI system development processes inherently prioritize security throughout the deployment pipeline.
Application Security Engineer
An Application Security Engineer focuses on integrating security throughout the software development lifecycle for applications. With the increasing integration of AI, securing these intelligent components within applications becomes paramount. This course delves into securing various AI implementations, including large language models and Retrieval Augmented Generation, which are often integral parts of modern applications. Learners are exposed to "secure by design" principles and critical considerations to reduce risk in organizational AI system development processes. An Application Security Engineer may find this course helpful in safeguarding AI technologies in production environments, addressing concerns like insecure plugin design and sensitive information disclosure within AI-driven applications. This particular course helps bridge the gap between traditional application security and the emerging complexities of securing generative AI components embedded within software products.
Security Risk Analyst
A Security Risk Analyst identifies, assesses, and prioritizes potential security risks to an organization's assets and operations, then recommends strategies for mitigation. The rapid adoption of generative AI introduces new and complex risk vectors that require detailed analysis. The Securing Generative AI course provides a comprehensive exploration of generative AI security, covering a wide array of AI threats such as prompt injection, training data poisoning, and aggressive agency. Learners are introduced to critical considerations and mitigations to reduce the overall risk in organizational AI system development processes. This course helps a Security Risk Analyst develop the practical knowledge of industry frameworks and best practices to safeguard AI technologies, enabling them to accurately assess and articulate risks specific to large language models and RAG implementations, and formulate effective risk management strategies.
Governance Risk and Compliance Analyst
A Governance Risk and Compliance Analyst ensures an organization adheres to internal policies, industry standards, and regulatory requirements related to security and data protection. As AI systems become more prevalent, ensuring their compliance and ethical operation is crucial. The Securing Generative AI course emphasizes "secure by design" principles and provides practical knowledge of industry frameworks and best practices to safeguard AI technologies. Learners are introduced to critical considerations and mitigations to reduce overall risk in organizational AI system development processes, which directly contributes to compliance efforts. This course can help a Governance Risk and Compliance Analyst understand the specific security requirements and potential vulnerabilities in large language models and RAG implementations, enabling them to develop and audit AI-related policies and ensure adherence to security outcomes and radical transparency.
Technical Program Manager Artificial Intelligence
A Technical Program Manager Artificial Intelligence oversees complex AI projects, coordinating across engineering, product, and research teams to ensure successful project delivery. This role requires understanding technical nuances, including security implications, to effectively manage risks and define robust project requirements. The Securing Generative AI course emphasizes "secure by design" principles and highlights building organizational structures that prioritize security, which is directly relevant for a manager guiding the entire AI system development process. Learners are introduced to various AI threats and comprehensive mitigation strategies for large language models and RAG implementations, equipping them to anticipate and proactively address potential security roadblocks. This course helps a Technical Program Manager Artificial Intelligence ensure security outcomes are integrated into project planning and execution from initial design to production deployment, ultimately facilitating the delivery of more resilient and trustworthy AI products.
Cloud Security Architect
A Cloud Security Architect designs and builds secure cloud environments, ensuring that all deployed applications and data comply with security best practices and regulatory requirements. Given that many generative AI systems, including large language models and RAG implementations, are deployed in cloud infrastructure, securing these environments is a paramount concern. The Securing Generative AI course provides insights into safeguarding AI technologies in production environments, which frequently reside in the cloud. Learners are introduced to critical considerations and mitigations to reduce risk in organizational AI system development processes. This course can help a Cloud Security Architect as it covers topics like securing vector databases, which are often cloud-native components, and understanding supply chain vulnerabilities applicable to cloud-hosted AI services. Building security-prioritizing structures for AI deployments aligns with the architectural duties of this role.
Software Engineer (Artificial Intelligence)
A Software Engineer Artificial Intelligence develops and deploys AI-powered applications and systems, often working with large language models, machine learning frameworks, and data pipelines. As the course focuses on "deployment and development" of AI implementations, understanding security from the outset is vital for building robust and trustworthy systems. The Securing Generative AI course introduces "secure by design" principles, which are essential for any engineer building AI solutions. Learners gain practical knowledge of securing various components, including orchestration libraries such as LangChain and LlamaIndex, and vector databases. This course can help a Software Engineer Artificial Intelligence to proactively embed security measures, addressing potential threats like insecure output handling and sensitive information disclosure during the development phase, ensuring the integrity and safety of their AI creations throughout their lifecycle.
Data Privacy Engineer
A Data Privacy Engineer focuses on designing and implementing systems and processes to protect sensitive information, ensuring compliance with data protection regulations. With generative AI models handling vast amounts of data, preventing sensitive information disclosure is a critical concern for this role. The Securing Generative AI course specifically discusses sensitive information disclosure as an AI threat and covers securing vector databases, which often store proprietary or personal data used by RAG implementations. Learners are introduced to critical considerations and mitigations to reduce the overall risk in organizational AI system development processes. This course can help a Data Privacy Engineer in understanding how to implement robust security measures around the data pipelines and outputs of large language models. This course helps strengthen the foundation for ensuring data confidentiality within emerging AI architectures.
Incident Response Analyst
An Incident Response Analyst detects, analyzes, and responds to cybersecurity incidents, containing breaches and restoring systems to normal operation. With the increasing adoption of generative AI, understanding the unique attack vectors and vulnerabilities within these systems becomes crucial for effective incident handling. The Securing Generative AI course introduces various AI threats, including prompt injection, model denial of service, and model theft. Learners gain knowledge of critical considerations and mitigations, which can inform the forensic analysis and containment strategies during an AI-related incident. This course may be useful for an Incident Response Analyst as it provides a foundation for identifying indicators of compromise specific to large language models and RAG implementations. Understanding how these systems are exploited can significantly enhance the ability to respond swiftly and efficiently to security breaches involving sophisticated AI technologies.
Research Scientist, Artificial Intelligence
A Research Scientist Artificial Intelligence conducts fundamental and applied research to advance the state of the art in AI, developing new models, algorithms, and techniques. While primarily focused on innovation, understanding the security implications and potential vulnerabilities of new AI models is increasingly crucial for responsible AI development. The Securing Generative AI course introduces a wide array of AI threats, including training data poisoning, model theft, and overreliance, which are critical considerations for researchers developing and evaluating new models. For a Research Scientist Artificial Intelligence, this course may be useful in integrating security considerations into the research design phase, promoting a more holistic approach to AI development. It helps build awareness of industry frameworks and best practices, enabling the creation of inherently more secure and robust artificial intelligence technologies from their inception.

Reading list

We've selected 50 books that we think will supplement your learning. Use these to develop background knowledge, enrich your coursework, and gain a deeper understanding of the topics covered in Securing Generative AI.
Co-authored by LangChain's Founding Engineer Nuno Campos and AI pioneer Mayo Oshin, O'Reilly's flagship guide covers modern LCEL (LangChain Expression Language), stateful multi-agent orchestration with LangGraph, production RAG, and LangSmith evaluation.
Is specifically focused on LLM security, including prompt injection as a key vulnerability. It offers a comprehensive guide to understanding, exploiting, and defending against attacks on LLMs, providing practical examples and techniques. This is essential reading for those looking to deepen their understanding of prompt injection from a technical perspective.
Guiding developers through implementing a GPT-style decoder in pure PyTorch, Sebastian Raschka demystifies tokenization, multi-head self-attention, pretraining loops, and weight loading. Subsequent chapters walk step-by-step through classification fine-tuning and instruction alignment.
This practical guide focuses exclusively on LLM security challenges and vulnerabilities, including a dedicated chapter on prompt injection. It draws on the collective wisdom from the creation of the OWASP Top 10 for LLMs, offering real-world guidance and strategies for developers and security teams. is highly relevant for professionals working directly with LLMs.
Pairing intuitive architectural illustrations with runnable code, Jay Alammar and Maarten Grootendorst explain transformer internals, semantic search embeddings, retrieval-augmented generation, and parameter-efficient fine-tuning. Engineers rapidly build mental models of how tokens flow through attention layers.
Ben Auffarth walks Python engineers through building production LLM applications with LangChain, covering prompt templates, vector stores, retrieval-augmented generation, custom tools, and deployment patterns.
This guide focuses on securing LLM applications, covering vulnerabilities and threats relevant to developers. It offers practical tools and techniques, including real-world case studies, to help safeguard LLM projects. is highly relevant for those involved in the development and deployment of LLM-based systems.
Authored by core Hugging Face engineers Lewis Tunstall, Leandro von Werra, and Thomas Wolf, this O'Reilly classic teaches encoder, decoder, and sequence-to-sequence Transformer architectures. Practical recipes cover cross-lingual transfer, model compression, and zero-shot inference.
Productionizing Retrieval-Augmented Generation requires advanced ingestion pipelines, hybrid search, query expansion, reranking, and RAG evaluation metrics. Practical architectural blueprints show how LlamaIndex fits into end-to-end LLM systems.
Provides a comprehensive overview of AI and ML fundamentals, with a strong emphasis on security, ethics, and privacy. It covers common threats, vulnerabilities, and attack vectors, including those relevant to LLMs, making it highly valuable for gaining a broad understanding of the landscape surrounding prompt injection. It useful reference tool for professionals and students alike.
This comprehensive guide is dedicated to the security of Large Language Models (LLMs), covering strategies, best practices, and future trends. It delves into the threat landscape and vulnerabilities specific to LLMs, making it highly relevant for understanding prompt injection in depth.
Chip Huyen provides the rigorous architectural foundation every LangChain developer needs: designing RAG pipelines, hybrid vector/BM25 retrieval, agentic tool execution, latency/cost optimization, and LLM-as-a-judge evaluation.
LangChain widely used framework for developing applications powered by language models, including RAG systems and AI agents. offers a practical guide to using LangChain and LangGraph, providing hands-on experience with implementing RAG and related concepts.
Alto's 2024 practitioner handbook walks through end-to-end RAG architectures, document chunking, embedding models, and production evaluation loops directly applicable to building LlamaIndex-powered applications.
Timely resource that explores the intersection of generative AI, cybersecurity, and ethics. It provides insights into the security risks and ethical considerations of generative AI technologies, which are directly relevant to understanding the broader implications of prompt injection. It is suitable for a wide audience, including leaders and professionals.
Dan Jurafsky and James H. Martin's foundational computational linguistics textbook establishes the mathematical bedrock of n-gram probabilities, perplexity, tokenization, information retrieval, and sequence modeling that every large language model researcher builds upon.
Published in 2025, Newton's focused reference pairs retrieval-augmented generation design patterns with vector indexing strategies, clarifying how LlamaIndex query engines ground LLM responses in private corpora.
Explores the challenge of aligning AI systems with human values. While not directly about prompt injection techniques, it provides essential background on the fundamental problem of controlling AI behavior, which is at the heart of prompt injection vulnerabilities. It's a must-read for understanding the broader implications of AI safety.
Ziegler's 2024 O'Reilly guide unpacks structured prompting, context-window budgeting, and few-shot synthesis techniques that govern how LlamaIndex response synthesizers turn retrieved nodes into accurate answers.
Updated in its second edition by Olivier Caelen and Marie-Alice Blete, this O'Reilly guide shows how to orchestrate OpenAI function calling, embeddings, and multi-step chains using the LangChain framework.
Goodfellow, Bengio, and Courville's canonical MIT Press reference provides the mathematical foundations of distributed representations, sequence-to-sequence attention mechanisms, optimization algorithms, and scaling laws underlying modern foundation models.
Written by a leading AI researcher, this book examines the risks of advanced AI and how to ensure it remains beneficial to humanity. It provides a foundational understanding of the control problem in AI, which is highly relevant to the challenges posed by prompt injection and other adversarial attacks aimed at manipulating AI behavior.
Large Language Models are a core component of RAG. provides a practical, hands-on guide to understanding and working with LLMs, covering transformers, tokenizers, and semantic search. It offers essential background knowledge for anyone building RAG systems and valuable reference for practitioners.
Denis Rothman bridges Hugging Face transformers, vision-language models, and retrieval-augmented generation in Python, walking practitioners through attention mechanics, fine-tuning open-weight models, and multimodal deployment.
Concise and hands-on, Mehul Gupta provides rapid code examples for chaining prompts, parsing structured JSON outputs, querying CSV/SQL data sources, and wrapping local Ollama/HuggingFace models.
This project-based book helps solidify understanding of Large Language Models and their applications, including the use of Vector Databases and LangChain, which are highly relevant to RAG implementation. It provides practical experience through building various LLM projects.
For those wanting a deep understanding of the generation component in RAG, this book guides you through building an LLM from scratch. It covers the internal workings, limitations, and customization methods. While challenging, it provides a solid foundation in LLM architecture and training.
Investigating scaling laws, tokenization quirks, out-of-distribution generalization, and inference-time compute strategies, Suhas Pai bridges LLM research insights with pragmatic application architecture and safety evaluation.
Aimed at beginners, this book provides a comprehensive roadmap to understanding Retrieval Augmented Generation (RAG) technology. It covers core principles, architecture, training processes, and real-world applications. This valuable resource for those new to RAG looking to gain foundational knowledge.
Explores vector databases and their role in modern data analytics, including applications in AI and generative AI. Understanding vector databases is fundamental to building efficient RAG systems. This book provides a good overview of this essential technology.
This guide for cybersecurity professionals covers adversarial AI attacks and defense strategies. Understanding adversarial attacks is crucial for comprehending prompt injection, as it form of adversarial manipulation. The book provides insights into threat modeling and securing AI systems.
More than half of James Phoenix and Mike Taylor's O'Reilly handbook is written directly in Python with LangChain, demonstrating few-shot selectors, Chain-of-Thought pipelines, Pydantic output parsers, and LangSmith tracing.
Drawing on AWS distributed training experience, Emily Webber and Andrea Olgiati demystify dataset curation, tokenizer design, and multi-GPU pretraining loops for teams building custom foundation models from raw text.
Microsoft AI engineer Valentina Alto guides developers through LangChain orchestration patterns, vector store indexing strategies, autonomous agent memory, and enterprise safety guardrails.
Authored by the creator of Keras, this book provides an accessible introduction to deep learning with Python. The third edition includes updated content on generative AI and large language models, making it highly relevant for understanding the generation aspect of RAG. It helps solidify understanding through practical examples.
Clear visual derivations of dot-product self-attention, residual streams, layer normalization, and contrastive pretraining make Simon Prince's MIT Press textbook an ideal mathematical companion for students studying Transformer architectures.
David S Richard demonstrates how to pair LangChain with Google Cloud services including Vertex AI and Cloud Run, offering practical blueprints for scalable, enterprise-ready generative AI pipelines.
Focuses specifically on applying deep learning techniques to natural language processing problems. It delves into advanced concepts and models relevant to the generative component of RAG systems. It is suitable for those looking to deepen their understanding of the neural network architectures used in NLP.
Focusing on application-layer orchestration, Ben Auffarth walks Python developers through tool-calling agents, structured output parsing, conversational memory, and vector-store retrieval chains. Practical guardrails and evaluation callbacks help move LLM prototypes toward production.
James Karanja Maina focuses on stateful, cyclic multi-agent workflows using LangGraph, teaching developers how to orchestrate branching logic, human-in-the-loop checkpoints, and persistent agent memory.
Balancing proprietary APIs and open-weight models, Sinan Ozdemir illustrates prompt engineering, semantic chunking, embedding clustering, and custom fine-tuning workflows. Concise case studies show how to align model capabilities with enterprise latency and cost constraints.
Considered a classic in the field, this book provides a broad and deep understanding of Natural Language Processing. While not specifically about RAG, it lays crucial groundwork in areas like text processing, language models (earlier forms), and information extraction, which are prerequisites for understanding RAG. It is commonly used as a textbook in academic institutions.
Inverted indexes, TF-IDF weighting, BM25 probabilistic ranking, and precision/recall evaluation metrics remain the mathematical bedrock of hybrid retrieval engines in LlamaIndex.
Save
Provides a comprehensive overview of vector GIS analysis, covering topics such as data models, spatial analysis techniques, and visualization. It valuable resource for GIS professionals and researchers.
Provides a comprehensive overview of red teaming techniques and tactics, including how to use them to evaluate the security of an organization.

Share

Help others find this course page by sharing it with your friends and followers:

Similar courses

Here are nine courses similar to Securing Generative AI.

Coursera Buyer Protection & Guarantee

Verified partner policies & purchase protections

Accredited University Partner
Free Trial & Audit

7-Day Free Trial & Free Audit

7-day free trial on monthly Coursera Plus subscriptions; free audit option available on most individual courses to explore video lectures and readings at no charge.

Return & Refund Policy

14-Day Refund Guarantee

14-day refund window on annual Coursera Plus subscriptions and single-course purchases prior to certificate issuance. Monthly subscriptions do not offer refunds once billed.

Buyer Protections

Verified Academic & Industry Partner

Accredited university and industry certifications (e.g. Stanford, Google, IBM, Yale), secure SSL payment processing, and verifiable digital certificate links.

Key Purchase Protections
14-day refund window on annual Coursera Plus and single courses
7-day free trial on monthly Coursera Plus subscriptions
Free course audit option for lectures and readings
Verified certificates from top universities & companies
OpenCourser Verified Partner Policy details last reviewed: 2026-10-22
Our mission

OpenCourser helps millions of learners each year. People visit us to learn workspace skills, ace their exams, and nurture their curiosity.

Our extensive catalog contains over 50,000 courses and twice as many books. Browse by search, by topic, or even by career interests. We'll match you to the right resources quickly.

Find this site helpful? Tell a friend about us.

Affiliate disclosure

We're supported by our community of learners. When you purchase or subscribe to courses and programs or purchase books, we may earn a commission from our partners.

Your purchases help us maintain our catalog and keep our servers humming without ads.

Thank you for supporting OpenCourser.

© 2016 - 2025 OpenCourser