AI research
Machine learning engineers and researchers. Language and speech processing, computer vision and model evaluation. A doctorate is welcome but not required.
Careers
Magic LEMP brings together researchers and engineers in Arcueil, France. We design AI systems for public institutions, defence organisations and businesses in regulated sectors. We also develop our own software.
We do not always have an open position, but we welcome conversations with people whose work impresses us.
The same team follows a project from beginning to end. They meet the people who will use the tool, test ideas on their data, build the software and support its deployment.
Our projects involve sensitive data. Much of the work is about running models where the data needs to stay: on a client’s infrastructure, with their access rules and available hardware. This is where the question of technological dependence becomes practical. We explore it in our thesis.
We also develop our own software, Dactylo, Djinny, Steno and Murmures. Each grew out of a specific assignment and continues to evolve with the teams who use it.
We hire as we meet people. These are the skills we regularly need.
Machine learning engineers and researchers. Language and speech processing, computer vision and model evaluation. A doctorate is welcome but not required.
Back-end and front-end engineers. Integration with our clients’ tools, with attention to code that works reliably in production.
Engineers working on on-site deployment, GPU servers, data isolation and monitoring.
AI consultants and project managers. Lead assessments, interview business teams and turn a need into trials with measurable results.
We are interested in less conventional backgrounds too. Our projects involve law, healthcare, industry and public debate. Someone who knows a profession from the inside and can write code often brings a great deal to the team.
We also welcome internships, work-study placements and industry-based PhDs. Tell us about your subject and the requirements of your programme.
We value people who measure before drawing conclusions. Comparing a model with a simple rule, keeping a clean test set and recognising when a trial has failed are part of our everyday work. They also help us tell a client when AI is not the right answer.
Sharing knowledge matters to us. We train our clients’ teams, and our president has taught at university for eight years. A tool is not finished if we cannot explain it to the person who will use it.
Founded in 2018 by two PhD graduates, Magic LEMP maintains close ties with research, particularly around Paris-Saclay. Everyone can see the effect of their work and take on responsibility early.
Write to us
Email us with “Candidature” as the subject, even if there is no open position.
Tell us briefly what interests you about working here. Include whatever best shows your work: a code repository, a paper, a dissertation or a project you have led.
A CV is enough. A cover letter is optional.
If your background matches our needs, we will get in touch to get to know you and discuss the projects we could work on together.