Coding agents
My method: specifications, a project harness, working scenarios and limits.
Artificial intelligence in software development and applications.
This section brings together my working method with coding agents, an AI application project and articles on how to assess quality, feasibility and security.
AI for software development
I use coding agents in projects, starting from specifications and checks. These pages describe the method and verification criteria.
Quality and coding agents
Code verification, maintainability and the constraints of legacy systems.
AI in applications
The RAG project described here is under development and testing. The other articles cover decisions to consider when integrating a model into an application.
AI applications
A RAG project for querying business data with precise results, still under development and testing.
AI feasibility
Data, initial preparation and service sustainability; technical answer quality is a separate assessment.
AI model selection
Task requirements, total costs, model lifecycles and migrations.
AI security
Authorisation, tool access, action boundaries and checks.