Projects
Sixteen projects, each described in depth and all without product or client names: products and integrations, sites and publishing pipelines, data and internal systems, and outreach platforms. Every case shows the problem, the architecture, what was delivered and the stack, with its real status and no invented metrics.
Sixteen projects in four groups, all described without product or client names. Together they cover SaaS products, e-commerce and billing integrations, public sites with their own data and APIs, AI pipelines checked by rule-based layers, internal analytics and compliant email outreach.
Products and integrations:
Sites and publishing:
Data and internal systems: support conversation analytics, engineering analytics on Git and tracker data and a community listening engine.
Outreach platforms built for one software client on a shared core: startup discovery and outreach, agency partner sourcing, affiliate reactivation and affiliate recruitment. Each instance runs isolated, with its own database, dashboard and sending identity, while opt-outs recorded by one are honored by all.
A wall of logos tells you little about how a team builds. A case that shows the data model, the build gates that stop a bad deploy, how scores are computed and how the data pipeline runs tells a CTO or a business owner far more. So each case covers the challenge, the solution, the functionality, the integrations and the stack, and states only facts we can verify in the code or on the running system, with a date.
No case has screenshots. Each one shows the pipeline stages and the engineering choices instead. For the products that means tenant isolation, idempotent webhooks and queues; for the sites, sourced data models, scoring engines and build gates; for the data systems, read-only connectors and rule-based checks on what the model says; for the outreach platforms, idempotent daily runs, deduplication, shared suppression lists and role-based dashboards.
If you want to see how these patterns apply to your own project, the services overview lists what we build, and get a quote starts the conversation.
FAQ
We describe every project without product names, client names, domains, data sources or volumes. Several were built for clients whose product plans or outreach strategy are part of their competitive advantage, and we present the other projects the same way. What we can show is the architecture, the engineering decisions and the stack, which is what matters when you judge how we build.
Every project here works and has been run end to end: on real data, against a live account or in a development store. Whether a system goes into public use is the client's decision, so not all of them are public. Each case page states its exact status, such as live, running on staging, in pilot or built and tested, and the date it was last checked.
Those figures belong to whoever owns each product, and for products launched in 2026 they would say little. We publish what can be checked in the code and the data instead: the architecture, the scoring and data rules, what was delivered and the current status, with the date each case was last checked.
The sites run on Next.js, React, TypeScript and MDX, hosted on Vercel. The products and integrations use Laravel, NestJS or Remix with PostgreSQL, Redis queues and React interfaces. The data systems and outreach platforms run on Python with FastAPI, SQLAlchemy and PostgreSQL. Each case page lists its full stack.
No. All sixteen projects are shown without product names, client names or domains. In a call we can go through the data model, the pipeline logic and the build and deploy gates of any of them in more detail, which tells you more about how we work than a logo would.
A Swiss technology company in Lugano.