SINHVU.

/about

A studio organized as a system,
not as a roster of services.

We are a small operation by design. The leverage is not headcount — it is the structure we work from, and the systems we leave behind.

.01 / Worldview

A studio is an operating discipline.

Most studios sell deliverables. We treat the studio itself as a system — a way of working that produces systems, not surfaces.

That changes everything downstream: the brief we accept, the questions we ask, the artifacts we hand over, and the moment we consider the work done.

.02 / Belief

Three commitments that shape every engagement.

  1. B.01

    The system is the deliverable.

    A finished project is not a file. It is a working system in your team's hands, with the documentation and rules needed to extend it.

  2. B.02

    AI is a co-operator, not a feature.

    We don't sprinkle AI on top of products. We design the points where it operates, the points where it must hand off, and the metrics that keep it honest.

  3. B.03

    Time is the real test.

    A system that looks good at handover but degrades at month four is a failure. We design for the third quarter of operation, not the first week of launch.

.03 / Method

Four movements, in order.

Diagnose. Design. Compose. Hand over. The method is consistent across every kind of system we build.

  1. M.01

    Diagnose the operating problem. Locate the actual constraint — workflow, team or decision — that the system will support every week.

  2. M.02

    Design the structure, not the surface. Tokens, taxonomies, decision rules, role boundaries. The visible layer is the last thing we draw.

  3. M.03

    Compose with AI as a co-operator. Place AI where it earns its place. Give it guardrails. Evaluate it like any other operator.

  4. M.04

    Hand over a system that runs. Documentation, training, a 30-day operating window — then your team owns it.

.04 / System

The four domains, held by a single core.

Design, AI, Product and Knowledge are not separate practices we offer. They are the four domains of one operating system — and the work is in how they hold each other in shape.

A diagram showing four domains — Design, Product, Knowledge, AI — connected to a central operating system core. Surface and Structure label the vertical axis; Input and Output label the horizontal.
Fig. 01 — Sinh Vũ operating system, schematic.
.05 / Organization

A small core, a wide network of operators.

We are deliberately under-staffed. The work scales through the system we operate from, not by adding bodies.

Core
An agency model: 30+ generalists, billable hours, layered management.
A small senior core that owns every project — system architects, design leads, AI operators.
Network
Sub-contracted to anonymous freelancers managed by a PM.
A trusted network of specialists brought in by name, accountable to the system, not to a brief.
Capacity
As many engagements as the pipeline can hold.
Three or four engagements at a time. We say no when we are full.

.06 / Culture

We work slowly on the things that matter, and quickly on the things that don't. The discipline is knowing the difference.

Internally, we run on the same systems we sell. Our design tokens, our internal AI workflows, our knowledge base — they are not products we built once and forgot. They are the operating environment we live in.

That means our engagements are field tests as much as deliveries. If a system we build for you is not the kind of system we'd run ourselves, we won't ship it.

.07 / Invitation

If your team is ready to build for time, we'd like to hear what you're operating.

Begin a conversation →