AI Assisted Secure Software Development

This immersive three-day course introduces software developers to the transformative role of artificial intelligence in modern development workflows. Participants will explore the landscape of AI-assisted coding tools, including GitHub Copilot, Amazon Q, and Tabnine, while learning prompt engineering strategies and security implications. Through hands-on labs, attendees will build an intelligent application using foundation models, deploy machine learning models locally and in the cloud, and develop red team techniques to understand and defend against common AI attacks. The course equips learners with future-ready skills essential for enhancing productivity, building smarter applications, and responsibly adopting AI in development environments.

By the end of this course, learners will be able to:

• Understand key AI concepts and their relevance to software development
• Employ prompt engineering techniques to interact effectively with language models
• Use AI-assisted development tools to streamline and accelerate coding workflows
• Build and enhance applications with AI-driven features and services
• Train and deploy machine learning models both locally and in cloud environments
• Identify and mitigate common AI attack vectors including prompt injection and model jailbreaks
• Apply AI securely in production-ready software development processes

Participants should have:

• A solid understanding of software development principles and practices
• Experience using integrated development environments (IDEs) and version control systems
• Familiarity with cloud environments (Azure, AWS, or GCP) is beneficial

Target audience

This course is ideal for:

• Professional software developers and technical leads exploring AI integration
• DevOps and security engineers seeking to understand AI risks and defences
• Development teams aiming to increase productivity and build intelligent applications
• Organisations adopting AI-driven workflows and tools

Day 1 – AI foundations and productivity tools

AI in software development

• Industry impact and emerging trends
• Enhancing productivity with AI
• Overview of AI-assisted development workflows

The AI landscape

• GitHub Copilot: features, use cases, and demos
• Amazon Q Developer: generating and securing code
• Tabnine: privacy, team collaboration, and coding efficiency
• Cloud-based lab setup and introductory exercises

Prompt engineering fundamentals

• Core techniques and strategies
• SudoLang for structured prompting
• Handling limitations and hallucinations in AI models
• Guided translation of natural language to structured prompts

Day 2 – Building intelligent applications

Advanced coding with Copilot

• Autocomplete, unit testing, boilerplate generation
• Integrated prompt engineering and Copilot Chat
• Real-world exercises in unfamiliar environments and frameworks

AI tool comparison and best practices

• ChatGPT in IDEs
• Feature and performance comparisons
• Choosing the right AI assistant

Application development with AI integration

• Fresh Cart project overview
• Integrating LLM-based chatbots
• Prompt routing and orchestration
• Semantic search and state mutation
• Logging and self-correction for AI operations

Security in AI systems

• Prompt injection: risks and red team exercises
• Model jailbreaks and defences
• Prompt extraction techniques
• Defensive strategies using ReBuff, Llama Guard, and Lakera

Day 3 – Working with models and deployment

Foundation models and inference

• Open-source models: Llama 3, Stable Diffusion
• Running models locally and in containers
• Hugging Face: transformers, pipelines, AutoTrain, and inference endpoints

Cloud-based AI deployments

• Microsoft Azure: OpenAI Studio and hosting
• Google Cloud: Vertex AI and Model Garden
• Amazon AI: Bedrock and machine learning toolkits
• Replicate and Cog for containerised model deployment

Machine learning essentials

• TensorFlow and Keras workflows
• Model training, tuning, and deployment
• Serving models using Docker

Advanced AI concepts

• Fine-tuning models in OpenAI, Vertex AI, and AWS
• Embedding techniques and similarity search
• Retrieval Augmented Generation (RAG) applications
• Research tools and embedding flows in Fresh Cart

Exams and assessments

There are no formal exams in this course. Learning is reinforced through practical exercises, red team simulations, and guided projects. Interactive labs provide a hands-on approach to mastering AI tools and methodologies. Follow-on courses we recommend Certified AI Security Engineer.

Hands-on learning

This course includes:

• Interactive labs using Copilot, Hugging Face, and cloud services
• Guided exercises for secure AI application development
• Red team scenarios simulating real-world AI threats
• End-to-end development of an AI-enhanced application
• Continuous instructor support throughout labs and activities