Work
AI systems, taken into production.
Each of these is structured the same way: the problem, the approach, and the result. Marked as representative where the underlying work is private or anonymised.
- MCP
- Secrets / Bitwarden
- Auth & scoping
- Multi-agent
- Orchestration
- Reliability
- n8n
- LLM automation
- Structured output
- Product
- AWS Bedrock
- Alfred
Have an AI system you need to make production-ready?
Whether it's an agent that needs to run unattended, an automation that has to be dependable, or AI features inside an existing product — let's talk about shipping it properly.