INSIGHTS
Thinking from the team
Perspectives on building, modernising, and operating AI systems — drawn from client work on AWS and Azure.
Choosing between Amazon Bedrock and Azure OpenAI for enterprise AI
Both platforms offer the same tier of leading AI models — so the decision usually comes down to where your data lives, what you already pay for, and what your team knows how to run. A practical framework for making the call.
READ ARTICLE →Multi-agent architectures that actually survive production
The hardest decision in an agentic system is not which framework to use — it is which parts should be an AI agent and which should be plain code. Lessons from an IoT monitoring stack built on AWS Step Functions, the Strands SDK, and Amazon Bedrock.
READ ARTICLE →Using LLMs to reverse-engineer the legacy system nobody documents
AI can map an undocumented legacy system in weeks instead of months — if you treat it as an analyst proposing theories, not an oracle. What it gets right, what it gets wrong, and how to check its work.
READ ARTICLE →The real cost of running AI features — and how to control it
AI features that are cheap to prototype can get expensive fast in production. Where the money actually goes, and the practical levers — caching, model routing, prompt discipline — that keep the bill under control.
READ ARTICLE →Want to talk it through instead?
If any of these topics touch a challenge you're facing, we're happy to share what we've learned directly.