RAG Beyond the Demo
Why a working retrieval demonstration is only the beginning, and what software engineers should consider when designing a production-oriented RAG system.
Technical Writing
Practical perspectives on production AI, software architecture and the engineering principles that help systems remain reliable, maintainable and designed to evolve.
These articles examine modern AI technologies through the wider lens of software engineering. The focus is not merely on making a demonstration work, but on understanding the decisions required to build complete software systems.
Topics include Retrieval-Augmented Generation, agentic systems, computer vision, architectural patterns, evaluation, observability and the long-term management of technical complexity.
Featured Articles
The article library is being developed alongside the practical engineering projects, live sessions and professional education programme.
An Engineering Perspective for Software Developers
A production RAG system involves much more than loading documents, generating embeddings, searching a vector database and sending context to an LLM.
This article introduces the wider engineering concerns: architecture, retrieval quality, evaluation, observability, reliability, security, cost and system evolution.
Engineering a Computer Vision Pipeline
A trained model is only one component of a production computer vision system.
This article will explore data preparation, preprocessing, inference, post-processing, evaluation, deployment, monitoring and the architecture surrounding the model.
Article in development.
AI-assisted development can generate code rapidly, but increased coding speed does not remove the need for clear system boundaries, modularity, testability and controlled change.
This article will examine why sound software architecture becomes more important—not less important—as development accelerates.
Article in development.
Browse by Topic
Articles will be organised around complementary areas of AI engineering, software architecture and long-term system design.
Designing, deploying and operating AI-enabled software systems beyond the experimental stage.
Retrieval quality, evaluation, architecture, observability and production concerns in RAG systems.
Tools, orchestration, workflows, responsibility boundaries and reliability in agent-based systems.
Complete vision pipelines involving data, inference, evaluation, deployment and monitoring.
System boundaries, interfaces, layers, deployment architecture and designing software to evolve.
Modularity, abstraction, coupling, cohesion and practical design decisions within application code.
Integration, distributed systems, workflows, orchestration and connecting AI capabilities to existing platforms.
Separation of concerns, design patterns, maintainability, documentation and disciplined technical decision-making.
Recent Writing
This section will grow into a chronological library of articles, technical notes and practical engineering investigations.
Why a working retrieval demonstration is only the beginning, and what software engineers should consider when designing a production-oriented RAG system.
A system-level examination of the components surrounding a computer vision model in a production pipeline.
How architecture, modularity and responsibility boundaries help AI systems accommodate changing models, providers and business requirements.
Writing Philosophy
Many technical tutorials explain how to reproduce a particular result using a framework or API. That can be useful, but it often leaves the deeper engineering questions unanswered.
My writing aims to examine why decisions matter, how components interact, what can fail and how systems can be designed for maintainability, observability and future change.
Software architecture is not about making today’s development harder. It is about making tomorrow’s change easier.
Open Engineering Resources
Where appropriate, articles will connect to source code, notebooks, architecture notes, diagrams, experiments and supporting repositories.
GitHub Repositories
Explore implementations, project structure, documentation, experiments and architectural decisions alongside the articles.
View GitHub profile →Practical RAG
A developing collection of practical RAG projects focused on software-engineering use cases rather than generic document chatbots.
Explore the repository →Live Education
Selected article topics will also be explored through live sessions, demonstrations and professional learning programmes.
View live sessions →Developing Library
The article library is intentionally being developed alongside real projects and teaching material. This allows the writing to emerge from practical investigation rather than abstract commentary alone.
Articles connect to problems encountered while building real systems.
Technologies are examined within the context of the complete system.
The emphasis is on principles that remain useful as tools evolve.
Continue the Conversation
Follow me on LinkedIn for new technical writing, project updates and announcements about future live sessions.