Independent AI & product engineering studio

AI products and systems, carefully built.

Comelse designs and engineers AI workflows, web and mobile products, and the integrations that hold them together. Founder-led, with a direct line to the person writing the code.

Selected work

Built around the way people actually work.

Three products from the portfolio, each shaped around a specific team and its daily routine.

01 · Retail platform · Knot Standard

OnDemand

Custom clothing, from the showroom to the factory. OnDemand brings product configuration, customer management and order workflows into one retail platform for Knot Standard.

Full OnDemand retail dashboard with new-order and client actions, Style Wall content, brand catalogues and garment selections. Account details anonymized.
Portfolio screenshot. Account and workspace identifiers anonymized; product interface preserved.
Our part
Long-term product engineering and integrations across ordering, CRM and operational workflows.
The system
Product configuration and client management connected to ordering, ERP, factory, logistics and payment integrations.
Disciplines
Web applications, retail workflows, systems integration.

Read the case study

02 · Generative imagery

TryOn Studio

A workspace for fashion teams to produce on-model product imagery. Brands define AI model personas, pair them with products and generate images for a whole catalogue in batches.

TryOn Studio asset library: a grid of AI-generated fashion images of models wearing catalogue products, with filters for product and AI model.
Portfolio screenshot. Account and brand details removed.
Our part
Product engineering across the generation workflow, the model and product libraries, and the asset management interface.
The system
Each batch crosses selected personas with selected products, queues the jobs and collects the results into a searchable asset library.
Disciplines
Generative imagery, job orchestration, web application.

03 · Sports performance

AthletIQ

A performance analysis platform that brings training load, wellness and strength data into one place, so staff can read a squad at a glance and plan the day around it.

AthletIQ GPS load dashboard with category summaries and a populated player table showing distance, player load, ACWR and status. Player identities anonymized.
Full dashboard screenshot. Player names, groups and dates anonymized; displayed metrics preserved.
Our part
Platform engineering, hardware data integrations and the analysis interface.
The system
Training load, wellness and performance information brought together for staff review, in a shared interface for the daily work of a performance team.
Disciplines
Data platform, device integrations, analytics interface.

Read the case study

04 · Computer vision

Mister Minit

Photo-based key identification for Clé Minit, Mister Minit’s mobile app. Computer vision helps turn a photograph into a matching key reference.

Official Clé Minit App Store images showing a key photographed in the app and the proposed replacement key.
Official Clé Minit promotional images, French interface. Source: MINIT FRANCE on the App Store.
Our part
Computer-vision engineering for photo-based key brand and model identification.
The system
A recognition workflow that connects a photograph of a key with a matching reference, supporting the duplication process.
Disciplines
Image recognition, machine learning, application integration.

Read the case study

Capabilities

From first build to everyday use.

One small team across the whole stack, so the model, the interface and the plumbing are designed together rather than handed between vendors.

  1. 01

    AI engineering

    Computer vision, language models and automated workflows, designed around a real task and wired into the tools your team already uses.

    • Image and document understanding
    • Assistants and internal tools with guardrails
    • Generative imagery pipelines
  2. 02

    Product development

    Web and mobile applications, taken from a first scoped version to a product with accounts, billing and the details that make it dependable.

    • Scoping and interface design
    • Web and mobile applications
    • APIs and back-office tools
  3. 03

    Integration & operation

    Connecting new software to your data, vendors and hardware, then keeping it healthy with monitoring, maintenance and steady improvement.

    • Third-party APIs and device data
    • Data pipelines and reporting
    • Hosting, monitoring and upkeep

Explore our capabilities

Studio

A direct line to the people building it.

Comelse is an independent, founder-led studio. Julien Renouard leads engineering and stays close to the work, from the first conversation to the final handover.

  1. 01

    Frame

    We start with the business problem, the data you have and the constraints you work within, then agree a scope and a first milestone in writing.

  2. 02

    Build

    Short cycles with working software early, reviewed together, so changes of direction stay cheap.

  3. 03

    Run

    We deploy, monitor and keep improving after launch, or hand over to your team with documentation.

Meet the studio

Have a system in mind?

Tell us what you want to build or fix, who uses it and where it hurts today. A first call is enough to see whether we are the right studio for it.