AI engineering system for regulated hardware
Engineers decide. Agents help them get there faster.
In regulated industries, the time to market of a complex product is measured in years: it has to be designed, documented and proven compliant. Koddex, a team of ten in Lille, exists to speed up every one of these steps and cut time to market by months, with AI agents that work beside the engineers.
What we build
From the BOM to the impact of every change, with AI agents in the team: the product in 70 seconds.
The team
Engineers, product people and developers who have worked inside hardware organisations, from robotics to aerospace.
Architect of the product and the vision. Helps rigorous organizations scale without sacrificing speed.
24 years building demanding tech products. Guardian of the platform's robustness and velocity.
10 years of product across startups and big tech. Turns technical complexity into operational value.
Code polyglot, obsessed with complex systems. Builds the deep infrastructure powering Koddex.
Self-taught developer obsessed with craft. From typed backend to pixel-perfect UI, ships fast and clean.
Backend engineer specialized in Rust. Builds the performant, safe foundations that run Koddex at scale.
Cloud-native and platform engineering specialist. Owns the infrastructure, CI/CD and observability that keep Koddex reliable, secure and scalable.
What we hold to
People first
Engineers keep the decisions and sign the changes. Agents take on the repetitive work, such as tracing links or checking coverage. We build for the people who are accountable for the product, and we hire people who care about it.
Rigor
We apply to our own work the rigor we build for our customers. In Koddex every figure has a source, every change has an owner and every decision keeps its reason.
Agility
We work in short cycles and each person owns their subject. We ship every week and talk directly with the engineers who use the product.
Innovation
We work on the problems that existing engineering tools leave open: keeping requirements, parts and tests consistent while the product changes every week, and giving AI agents a real share of the engineering work, within the rights of the team.
AI in engineering, on data you can trust
An agent is only as good as the context it reads. In most engineering teams that context is scattered across documents, spreadsheets and people's memory, so an AI can only guess.
Koddex starts by structuring and linking the requirements, BOM, system model, tests and decisions. Agents then take the tasks assigned to them like any teammate: they create and update items, comment, move workflows forward and hand work to other agents, with the permissions of the user they work for. Every action they take is logged and reviewed like any other change.
Where engineering is heading
In the next few years, every engineering team will work with agents beside them. In regulated and certified hardware, those agents will only be trusted if they act on structured data, within permissions, with every action traced.
Koddex is built on a secure, traced source of truth where every piece of data and every decision is versioned and verifiable. Engineers and agents collaborate on it the way developers already do with code: the agent executes, the engineer validates, and the structure keeps every action traced and within bounds.
- 1,000+
- engineers using Koddex
- 2024
- year founded
- €3M
- raised, seed stage
- 10
- people on the team
Backed by
