Skip to content
About

Seven people who would rather be rightthan fast.

SketchLabs is an independent product studio in Colombo. We design and build software for teams that cannot afford to get it wrong twice — usually because the software is load-bearing for something regulated, expensive or irreversible.

01 / What we work with

Enterprise platforms

Accredited or certified, and implementing them weekly rather than once a year.

  • Odoo
  • Microsoft Dynamics 365
  • SAP Business One
  • Salesforce
  • HubSpot
  • Shopify Plus
  • WooCommerce
  • Make / n8n / Zapier
Enterprise practice →

Models we build against

Chosen per task against an evaluation set, never by default.

  • OpenAI
  • Anthropic
  • Google
  • Meta (self-hosted)
  • Mistral
  • Open models (via a provider)
AI practice →

Everything else

Boring, well-understood tools. Novelty is a cost, and we spend it only where it actually differentiates the product.

  • TypeScript
  • React
  • Next.js
  • PostgreSQL
  • Kubernetes
  • Python
Years shipping
6+
Projects delivered
25+
People in the studio
7
Clients return
92%

02 / Technologies we use

PHPTypeScriptReactNext.jsNode.jsPythonPostgreSQLRedisTailwind CSSGraphQLDockerKubernetesOdooSAPSalesforceFigmaPlaywright

03 / What we believe

  1. 01

    Incremental over heroic

    We would rather migrate a system over nine months than rewrite it over one. Every engagement we have ever regretted started with a rewrite someone was very confident about.

  2. 02

    Measure before optimising

    Profiling beats intuition. We instrument first, fix what the data points at, and re-measure to prove it — including when that proves we were wrong.

  3. 03

    Accessible by construction

    Keyboard support, focus order and contrast are decided while we design, not audited at the end when they are expensive to change.

  4. 04

    Leave it maintainable

    The test is whether your team can extend the work without us. Handing over a black box is not a deliverable.

  5. 05

    Say the difficult thing early

    If a deadline is unrealistic or an idea will not work, we say so in week one rather than week nine. It is much cheaper.

  6. 06

    Boring where it counts

    Novelty is a cost. We spend it on the parts that differentiate the product and use well-worn tools everywhere else.

04 / The studio

05 / How we got here

  1. 2019

    Founded

    Three people, one client, and a strong opinion that migrations should not require a rewrite.

  2. 2021

    First design systems engagement

    A consolidation project for nine teams taught us more about adoption than any technical decision ever has.

  3. 2023

    Practice grew to seven

    We started turning down more work than we accepted, which remains the most useful thing we have done.

  4. 2024

    Practice dedicated to AI integration

    Because someone has to build the evaluation harnesses before the features ship.

  5. 2026

    Twenty-three projects delivered

    Still independent, still deliberately small, still measuring everything twice.

06 / Clients

  • MeridianFinancial services
  • FieldnoteHealthcare
  • NorthwindLogistics
  • AtlasInfrastructure
  • CadenceDeveloper tools
  • HeliostatSaaS
Work with us

We are hiring, and we take on projects.

If you are looking for a studio that will tell you when something will not work, get in touch. We are also always happy to meet people who want to join.