Workstream 1
One of the earliest realizations I made is that LLMs cannot be trusted to get architecture right, and cannot be trusted to be consistent. In this workstream we’re building a pipeline that enables feature requests to be reliably broken down into thin slices called capabilities, each with over 100 structured inputs to ensure the a backend feature is properly built.
learn moreWorkstream 2
Many tech organizations can struggle with building reliable front-ends. With this work we’re building a design-system that is governed, reviewable, and structured in such a way that any app features must reuse the same seam (style, behaviour, etc.) that has been signed off under the design system.
learn moreWorkstream 3
How do we make the work that’s being done by the Agentic harness more visible and human approvable through the use of stories and tasks? Starting with the stakeholder interview and appropriate routing of follow on questions and decision points based on subject matter.
learn more