IoT platforms
Connecting and securing device fleets: AWS IoT Core, secure device registration and provisioning, over-the-air updates, signed firmware, cellular or Wi-Fi telemetry, and the fleet administration tooling that has to come with it.
Product Development
Products that reach production, and stay there
We design and build platforms on AWS, from the first framing session to the run phase. Two people, code we stand behind, and an architecture built to be handed over.
Why hand a product to two people
Software products rarely die from a shortage of hands. They die from a pile of decisions made by people who were no longer around to answer for them. Staying two keeps the thread intact: whoever chose the data model is the one who will fix it six months later.
We take on projects where the difficulty is technical and where the person opposite knows what outcome they want. We do not take on projects where our role would be to fill a development capacity slot.
We have built IoT platforms connecting large consumer device fleets, an end-to-end encrypted health platform between practitioners and researchers, an industrial production management system, scientific data pipelines, and GPU infrastructure for generative AI. The detailed cases are at the bottom of this page.
What we build
Six kinds of work, all of them shipped to production rather than prototyped and abandoned.
Business applications and back-offices
The interfaces your teams live in every day: production management, fleet administration, after-sales support, order tracking. Unglamorous, and the difference between a platform that is used and one that is tolerated.
APIs and integrations
Documented REST APIs, webhooks, MCP servers to expose your data to agents, connectors into the systems you already run. GDPR compliance and data portability obligations such as those introduced by the Data Act.
Data and analytics
Ingestion and transformation pipelines, warehouses on AWS or Snowflake, and the reporting on top. We have taken scientific analyses that took days and brought them down to minutes.
Machine learning and generative AI
Predictive maintenance, document agents over your own technical references, scalable GPU infrastructure to run specialised models. We stay on the engineering side: getting it into production, and keeping cost, latency and reliability under control.
Application security
End-to-end encryption where the subject demands it, data isolation, authentication, auditability. When confidentiality rules out sending data to a third-party API, we run AI models locally on our own GPUs. On a health platform, none of that is optional.
How we work
Short cycles, deployed increments, and no phase where you cannot see progress.
Short framing phase
A few days to turn an intention into a scope, with a target architecture and an estimate. The output is written and useful even if you decide not to continue with us.
Incremental development
Regular deployed releases, not demos. You see the thing you are paying for running, early and often.
Production
Infrastructure as code, separate environments, observability, backups. Going to production is not an event at the end, it is a property of the project from the first week.
Handover or managed services
Either we hand you the product with its documentation, or we operate it for you. Both are legitimate, and the choice is yours.
Related case studies
All case studiesQuestions we get asked
There are two of you. How long does a product take?
A first scope running in production is usually measured in weeks, not quarters. We would rather ship a reduced scope that works than a complete specification that waits.
Who owns the code?
That is yours to decide. Most clients prefer full ownership from day one, in their own repository. We can also keep it on our side if that is what makes sense for the project. Either way, we retain no proprietary component to create a dependency.
What happens if you are no longer available?
That is the right question to ask a two-person firm. Our answers: the code is yours, the infrastructure is described as code, the documentation is written for a stranger, and the Cloud Partners collective can pick it up. We design to be replaceable.
Fixed price or time and materials?
Framing is fixed price. Development is usually by the day, because that is the honest model when scope moves, and we flag drift as soon as it appears rather than at the end.
Will you work alongside our own developers?
Often, and it is the best arrangement. We bring the architecture and the hardest parts, your team keeps ownership of the domain.
What technologies do you use?
We are deliberately conservative about structural choices: TypeScript, Node.js and React on the application side, Python where the data work calls for it, Terraform or CDK for infrastructure, PostgreSQL or DynamoDB depending on the actual access pattern. We do not introduce a technology because it is new, but because you will still be able to hire someone who knows it in three years.
Will you take over a project already under way?
Yes, and it is often a project in trouble. We start with a short assessment and tell you honestly what can be salvaged and what would cost less to rewrite. We have turned down takeovers where a rewrite was the only sensible answer, because pretending otherwise would only move the failure a few months out.
The other two things we do
AWS Consulting
We audit existing AWS estates, design architectures that hold, and bring costs back under control. Short, scoped engagements run by the two of us, based in Lyon.
AWS Hosting & Managed Services
Hosting on AWS, on-call, monitoring, application and infrastructure maintenance, and continuous cost control. We bill infrastructure in euros, and we answer when it rings.
Talk to us directly
Write to us directly. No salesperson, no qualification call. One of us two replies, within 24 hours.
contact@onescale.io
We reply within 24 hours, and it is one of us who replies.
Based in Lyon, working with clients in France, across Europe and internationally.