At Fretlink we love stability, attention to details, a bit of complexity and open-source. Our tech has many shapes and connections which we are presenting here. Most of those shapes are there to support our lorry transportation management system, our users and our internal team.
Handling truck-loads of business rules can be hard without strong machine guaranties. A compiler and a strong type system help us feel confident about our changes. It also ensures our technical modelling follows our team discussions. Additionally this makes our refactoring time as fun as our mario-karting (or Euro Truck simulating) time.
What is the best Relational Database Management System? We don't have an answer but we really love PostgreSQL as a software. It is powerful and we trust our data is well taken care of on the back of an Elephant. Also they have a nice set of tools and on top of that, a lovely community.
The team is still trying out new things and that's good! We have a web service and some internal tooling written in Rust. Not only it's a robust language offering good performances when needed, but it also makes our developers happy. That's a good beginning!
Ansible + Terraform
Part of our job is to pilot transportation demands from A → B. For infrastructure needs we prefer to let Terraform do the piloting job for us. It helps a lot to keep a shared, versioned and controlled state of our systems within the team.
Machine configuration is left in the hands of Ansible.
Have we said something about types already? Well, we love them. When we can our YAML files are authored through dhall. Not only it provides static checking of our configuration files, but it also removes the need for ad-hoc YAML templating systems.
Eating organic and local has its benefits. That's why we really enjoy hosting our main services on Clever-Cloud. We still don't like to keep all our eggs in the same bucket though, so we also have some resources on AWS, OVH and Hetzner.
Research & Development
Bibliographic sources are treasures of thoughts for our research. We search for the Juste prix (The Price Is Right), routing optimisation with geographic constraints and much more. We don't have a typical day, but bibliography is part of our daily routine. We remain full-fledged researchers by participating in scientific community through partnerships with universities and encourage high-level publications.
Paper & pencil
The set of disciplines we tackle is diverse. They cut across spatial econometrics, statistics, costing, stochastic optimisation, data analysis. The supply chain and road freight transport are in full evolution, so our perspectives and possibilities are numerous and still to be developed: stochastic control, reinforcement learning, optimal transport, risk management, demand prediction, demand planning... There is room for creativity!
Did we tell you we like PostgreSQL? Together with PostGIS it gets even better to work with geographical data. We use data as diverse as their complexity such as socio-economical data-sets, geographical routes, distances and points, etc…
R and Python
R like Research? At least for maths and statistics it helps a lot.
Both languages are the first steps between our paper thoughts and our final system running in production.
To write our technical reports, articles and presentations we can always count on the good'ol LaTeX engine.