Analysis of Legal Decisions Intern

Location: 64000 Pau, France

Development of a legal decision processing pipeline for information extraction and training of text generation models.

At Lipstip, a young innovative company created in March 2023, we combine expertise in artificial intelligence and intellectual property to help intellectual property experts generate better and faster protection for innovative European companies.

In concrete terms, we are developing a software suite incorporating artificial intelligence to defend trademarks, draft patent applications and facilitate innovation protection procedures.

Protecting innovations is essential. In Europe, protecting your innovations means 68% more revenue per employee for companies that own a patent or trademark, yet only 10% of companies protect themselves. So, by helping innovation protection professionals, we can contribute to the performance of all European companies.

We have initiated a research partnership with CY Tech and are in discussion with other French universities. We have been awarded the Startup Program by OVHcloud, which gives us access to unprecedented computing resources for the year 2023.

We would like to welcome a work placement student who is motivated to contribute to our patent application drafting tool, which is currently being finalized.

This subject, relating to automatic language processing, will provide an ideal framework for machine learning in the sense that a very large, relatively clean corpus will be accessible. In particular, the aim of the placement will be to develop a pipeline for processing legal decisions to extract conclusions and reasoning.

During this course, we will be asked to use lightweight, low-energy consumption solutions to approach the accuracy of costly, complex state-of-the-art solutions.

In addition to this, a crucial and original point will be to build an interpretable model whose results can be easily explained by one or more direct or indirect links with the data.

During this placement, you will be required to test various data pre-processing, and several models for extracting unnamed entities and relationships between entities. In addition, you will have the opportunity to use the extracted data to train an LLM-type model to generate texts whose content can be easily explained. Your main assets will be your fluency in algorithms and Python. Finally, during this placement you will be made aware of intellectual property rights and the implications of these rights for your professional practice.

We look forward to hearing from you about your background, your motivations, and our internship opportunity.

Apply with CV and cover letter on