Muselink
How do you architect a music collaboration platform where every technical decision must be derived from user behavior — across four technology layers — while maintaining a UX quality bar that rejects easier but inferior solutions?
Muselink is a personal venture I have been building since 2015. The platform requires seamless audio upload, waveform trimming, frequency-accurate visualisation and *
goal-based collaboration matching*.
The core architectural challenge is that none of these features can be designed in isolation. Each one shapes the technical constraints of the others. The upload flow defines the data model, the visualiser demands native rendering and the *
collaboration layer depends on both*.
This is not a design project with engineering support. It is *
product-driven system design* where every UX decision carries architectural consequences.
Demonstrated that a single product leader can architect and ship a multi-layer platform through constraint-driven direction
Proved that taste in product + architecture standards is the competitive moat, not raw engineering output
Established a repeatable model for AI-directed product development with quality governance
Built a foundation that can scale from solo venture to team-led platform
Product architecture is derived from user behaviour, not engineering defaults — a principal-level design practice
Architectural governance rules (logging-first, no global overlays, MVP-before-features) demonstrate staff-level systems thinking
Cross-domain orchestration across frontend, native, backend, and infrastructure proves ability to lead complex technical products
AI-directed development model shows the future of design leadership — defining problems and enforcing quality at scale
Design Leadership
Four capabilities observed during the Muselink build that define how I operate as a Product Systems Architect
Desirability Before Feasibility
I consistently optimise for experience quality first, then solve for the technical architecture to support it. The Muselink visualiser had to feel real — not a simplified amplitude animation. That meant moving computation to native iOS (Metal + Swift) and rejecting several technically simpler implementations that did not meet the experience standard. Most builders do the opposite: they optimise for engineering simplicity. Products that win — from Apple's hardware-software integration to Airbnb's experience-first redesign — start from experience quality. This is the approach I bring to every system I shape.