How we approach the work.
We publish case patterns, anonymised write-ups of how we approach a class of problem, instead of stamping client logos we don't have permission to use. Patterns are written by the engineer who did the work.
Method pattern
Designing AI agent workflows humans can trust
How we structure agentic workflows so the agent does the routine path, humans decide the exceptions, and everyone can audit what happened. Deterministic logic, structured outputs, evaluation harnesses, observability over prompts and tools.
Read the patternMethod pattern
Audit-trail and evidence design for regulated software
Evidence schema and audit-log architecture for software that needs to support a regulator question, an auditor request, or an internal investigation later. What gets logged, what gets retained, what gets indexed, who can read it, who cannot, and how long it stays around.
Read the patternMid-market, AU
Migrating a legacy .NET system to a cloud-native architecture
Phase-by-phase rebuild of an in-production .NET monolith into a cloud-native architecture with zero customer-visible downtime.
Read the patternFintech / AI software, AU
Building an AI-augmented operations platform for an Australian fintech
Multi-year build of a CRM, document workflow, and AI document-intelligence platform for an Australian fintech and AI software business. .NET, React, Azure, AI document classification and extraction with human-in-the-loop on consequential outputs.
Read the pattern