Source our solution. Drive your business.
We build the systems institutions run on — ERP, PLM, education platforms, indoor maps — and the data, AI and engineering capacity around them.
Fashion PLMRecords
- SKU
- Colorway
- Style
- Season & hierarchy
- BOM revision
- Material
- Review
- Audit history3
Cable Knit Crew
Season
AW26
Gender
Women
Product group
Knitwear
Target launch
1 Jul 2026
Merino Lambswool 2/28 Nm
status
Concept→In Development
style_phase
Proto→SMS
authority_bom_id
—→640
At the helm
You set the course. Our products hold it.
Five products on one wheel, and one architecture under all of them. Each takes the helm for its stretch of the voyage and loads the work it owns, so the ship reaches harbour with everything aboard.
Cast off
Work drifts without a helm.
A style waiting on its BOM, a booking nobody has received, a class list on paper. Each piece sets its own course until someone takes the helm.
Still adrift
- product_passport
- booking_no
- register scan
- capacity
- auction cycle
Traceability and Digital Product Passport for apparel
Fashion PLM
Digital Product Passport ready, because the data was always there.
Explore Fashion PLMLoaded at this light
- product_passport
- chain_of_custody
- season_code
- style_phase
- Explore Fabrivent
Loaded at this light
- booking_no
- return_supplier
- gate_pass_url
- storage.address
- Explore Shohojogi
Loaded at this light
- register scan
- class list
- term result
- progress report
Capability showcasing for garment factories
Manufacturer CMS
Show a European buyer what your factory can actually do.
Explore Manufacturer CMSLoaded at this light
- capacity
- machinery
- certification
- compliance
Recurring auctions, run without the coordination
Auction Platform
Set it up once. Auction every month.
Explore Auction PlatformLoaded at this light
- auction cycle
- bid ledger
- bidder verification
- winner selection
Harbour
Drive your business.
Five lights passed, every piece of work aboard, and the whole ship in harbour.
- engineers
- 10+
- Building sophisticated products end to end, not filling seats.
- products in production
- 5
- Fashion PLM, Fabrivent, Shohojogi, Manufacturer CMS and Auction Platform.
- services
- 5
- Map & Geospatial, Data & Analytics, AI Software & Agents, Custom Software and Engineering Talent.
- technologies
- 40+
- Cloud, data, backend and frontend — listed in full, not as logos.
The stack we build on
Products
Five products we own, run and sell.
Not client work. Our own systems, in production, on one architecture — which is why the PLM and the ERP agree on what a style is made of.
Fashion PLM
Digital Product Passport ready, because the data was always there.
Explore Fashion PLMRecords in the PLM
Cable Knit Crew
Nordwear · D24309 · Fog Grey
5400123456789
Shohojogi
Premium education software, with AI doing the paperwork.
Explore Shohojogi
Manufacturer CMS
Show a European buyer what your factory can actually do.
Explore Manufacturer CMSAuction Platform
Set it up once. Auction every month.
Explore Auction PlatformTraceability
One thread, from a shipment back to the fibre.
Pick a delivery in Fabrivent and walk it backwards — through the job, the SKU and the approved BOM — to a validated fibre composition in the PLM. Every step is a record in the shape the two systems store it.
- DeliveryDelivery #131Fabrivent
- BookingBook-2026-118Fabrivent
- JobJOB-214Fabrivent
Fabrivent hands off to the PLM
SKU5400123456789PLM- StyleD24309PLM
- BOM revisionD24309 Production BOM · rev 4PLM
- MaterialMAT-4471PLM
Why now
EU Digital Product Passport registry goes live
ESPR (EU) 2024/1781
Textile delegated acts expected
ESPR (EU) 2024/1781, Article 4
EU Forced Labour Regulation applies
Regulation (EU) 2024/3015
Fibre-level traceability mandatory
ESPR textile requirements
Services
Work that starts with your question.
Five practices, each built around a problem clients kept arriving with.
Map & Geospatial
“How does anyone find anything on a site this big?”
Google and Mapbox stop at the front door. We build site maps that go inside — pick a room, a shop or a sewing line and get a walking route that changes floors by lift, escalator or stair.
Includes
- 3D outdoor site maps for campuses, EPZs and industrial zones
- Indoor floor plans with vertical circulation
- Turn-by-turn wayfinding across buildings and floors
- Destination search across units, rooms and facilities
- Route planning, distance and travel-time calculation
What you get out of it
We own the data format, the viewer and the routing graph, so there is no tile provider in the loop and nothing to meter. A venue works offline, and the bill does not scale with your traffic.
Data & Analytics
“Why does every report give a different number?”
Scattered, unstructured and contradictory data turned into something a business can decide from — collected, cleaned, modelled and made available where people actually work.
Includes
- Data pipeline design and development
- ETL and ELT process automation
- Data cleaning, validation and transformation
- Data warehouse architecture and implementation
- Centralised data platform setup
What you get out of it
One place the numbers come from, with the lineage to show where each one originated. Disagreements become traceable rather than political.
AI Software & Agents
“Where does AI actually save us time here?”
Applications and agents that read your documents, answer from your data, and complete workflows in the systems you already run — built to be evaluated and monitored, not demoed once.
Includes
- Custom AI application development
- Agents for business workflows
- Retrieval-augmented generation over internal knowledge
- Document analysis and structured data extraction
- AI-assisted search across internal systems
What you get out of it
We start from the workflow, not the model. If a rule beats a model we will say so — and the things we ship come with a way to measure whether they are still working.
Custom Software
“Who can build this so it is still maintainable in three years?”
The platforms behind this site are our own: two Go services on Clean Architecture with role-based access, audit trails, generated API documentation and Kubernetes deployment. We build yours the same way.
Includes
- Product discovery and technical architecture
- Backend services in Go, Python, Java and Node
- Web applications in React, Next.js, Angular and Vue
- PostgreSQL and MySQL data modelling
- Role-based access control and audit logging
What you get out of it
Layered architecture, tests around the parts that matter, and documentation generated from the code rather than written beside it. You can hand the result to another team.
Engineering Talent
“We need senior capacity without an 18-month hiring cycle.”
Data scientists, data engineers, software and DevOps engineers who embed in your workflow, your stack and your standards — and are accountable for outcomes rather than hours.
Includes
- Data scientists and ML engineers
- Data and analytics engineers
- Backend, frontend and full-stack engineers
- DevOps and platform engineers
- Forward-deployed engineers for client-side delivery
What you get out of it
We embed into your workflow and communication rather than running a parallel process. Senior capability at rates that work for a growing company, without the overhead of headcount.
How we work
The same three steps, whatever the work.
- 1
Understand
The question you actually have, the data you actually hold, and who has to live with the answer. Usually a week, sometimes a call.
- 2
Build
A small first release in your stack, with the parts that matter tested and the API documented from the code. You see working software early, not slides.
- 3
Run
Monitoring, a handover your own team can operate from, and a written record of every decision. We are easy to keep and easy to leave.
How we build
We do not have a team page. We have the systems instead.
The two platforms this site is built around are ours, and they are built the same way we would build yours. Here is what that means in practice.
Clean Architecture, enforced by the layout
Ports define the contracts, application holds the use cases, framework holds HTTP, database and external services. Business logic does not import a web framework, so it can be tested without one.
Permissions on a resource:action model
Access is checked server-side against explicit permission codes, not inferred from a role name in a token. The UI gates to match, so a user is never offered a link the API will refuse.
Audit trails as a data structure
Changes are recorded with their before and after, not appended to a log file. That is what makes 'who changed this' a query rather than an investigation.
Generated API documentation
Swagger generated from the code, so the documentation cannot drift from the endpoint it describes.
Configuration outside the binary
Config is read from Consul at startup rather than baked in, so the same image runs in every environment.
Containers and Kubernetes
Multi-stage builds onto Alpine, deployed to Kubernetes. Deployment is a described process, not tribal knowledge.
The stack, in full
- Backend
- GoPythonJavaNode.jsRust
- Frontend
- ReactNext.jsAngularVueSvelteTailwindFlutter
- Data stores
- PostgreSQLMySQLRedisBigQuerySnowflakeRedshift
- Data platform
- AirflowdbtKafkaSparkFlinkAirbyteIcebergTrinoPinot
- Analytics
- SupersetMetabaseRedash
- Cloud
- GCPAWSAzureHetznerDigitalOceanAlibaba CloudHuawei Cloud
- Platform
- DockerKubernetesTerraformConsulNGINX
Questions
What people ask before the first call.
Every answer below is lifted from what the products and practices on this site actually do.
How does the PLM handle the Digital Product Passport?
Composition, origin, certification and end-of-life data are held per product and assembled into a passport. Because they are the same records production already works from, the passport is a view of live data rather than a document someone has to reconcile.
Do the PLM and Fabrivent actually share records?
Every SKU resolves back through colorway, style, BOM revision and material to a validated fibre composition — and forward into the ERP, so a shipment and a specification are answers to the same question.
Do your maps depend on Google or Mapbox?
We own the data format, the viewer and the routing graph, so there is no tile provider in the loop and nothing to meter. A venue works offline, and the bill does not scale with your traffic.
Will you tell us if we do not need AI?
We start from the workflow, not the model. If a rule beats a model we will say so — and the things we ship come with a way to measure whether they are still working.
Can another team take over what you build?
Layered architecture, tests around the parts that matter, and documentation generated from the code rather than written beside it. You can hand the result to another team.
Can we add engineers to our own team instead?
We embed into your workflow and communication rather than running a parallel process. Senior capability at rates that work for a growing company, without the overhead of headcount.
How does an engagement start?
The question you actually have, the data you actually hold, and who has to live with the answer. Usually a week, sometimes a call.
What happens to what I send through the contact form?
To reply to your enquiry and to keep a record of the conversation if it becomes a commercial relationship. That is the only purpose.
Get in touch
Tell us what your data cannot currently answer.
A short conversation is usually enough to tell whether this is a system problem or a process one. We will say which.
- [email protected]
- +880 1672 868711
- Replies within 24 hours