3D Database Internship
Pubblicato il 06-09-2026 - Liablix in Firenze
3D Database Internship
€400/month allowance
Bologna or Florence, Italy.
This is a hybrid role, but on-site presence is preferred.
Who we are
At Liablix, we are building AI that reconstructs road accidents from images of the vehicles involved, helping insurers determine whether a claim is fraudulent and assess liability. We are backed by leading international investment funds that strongly believe in our vision and support our rapid growth.
At the core of our technology is the 3D vehicle: an accurate digital replica of its real-world counterpart. These models are stored in a database we have built and continuously expand and update. We are looking for someone to help us build and maintain it, contributing directly to the development of our product.
Why this internship is different
You won’t simply be given a fixed list of tasks. You will be encouraged to understand where you can add the most value and bring forward your own ideas to improve our platform and organize the 3D vehicle database more effectively.
You will also work closely with our highly skilled tech team,
collaborating daily with experienced software developers. It’s a valuable opportunity to learn how a fast-growing technology startup builds and improves its product.
What you'll do
Expand our 3D mesh database by identifying the best models available and selecting those that meet our quality standards. Keep the database continuously updated by identifying and sourcing newly released vehicle models.
Determine how to organize the database as effectively as possible. With many variables to consider, you will help define its future structure and how it integrates with our platform.
Work closely with AI researchers and software developers to improve our workflows. This collaboration will be part of your daily work and offer continuous opportunities to learn from them.
Perform Quality Assurance on 3D meshes processed by our technology, evaluating the accuracy and reliability of AI-generated results.
Improve da