Company description
Meritis is a consulting firm specialised in the digital transformation of organisations, founded in 2007 by Sébastien Videment.
Initially based in Paris, it has expanded across the regions and now has a presence in France's largest cities: Sophia-Antipolis, Montpellier, Nantes, Bordeaux, Lyon, Aix-en-Provence, Lille, and even Lisbon in Portugal since 2023.
Our experts support clients from various sectors with all their digital transformation needs across many areas of expertise: Finance, Software Engineering, Cloud & Infrastructure, Data, Digital Transformation and Cybersecurity.
Guided by its values of excellence, humility, benevolence and closeness, the firm of more than 900 employees, which has been awarded 5 times in the Great Place To Work® rankings, is experiencing very strong growth and aims to reach €100M in revenue in 2024 and surpass the symbolic milestone of 1,000 employees.
We take pride in being close to our employees and supporting them individually, regardless of their role in the company. Certified by Great Place To Work since 2013, our approach to workplace well-being goes well beyond a simple label; our employees speak about it best: https://www.glassdoor.fr/Avis/Meritis-Avis-E1163008.htm
Position description
Navigacom is the infrastructure-specialised entity of Meritis (Enterprise Networks, Cybersecurity, Digital Workspace, Cloud & Infrastructure). We are developing an internal AI platform designed to accelerate and standardise the production of mission deliverables and to automate its internal processes. The internship will be carried out in a pair, within the team leading the project, in direct contact with consultant users. The scope is operational: the developments are intended to go into production, not remain prototypes.
Main responsibilities
The pair will be in charge of the platform end to end:
- Map the recurring deliverables and internal processes of the firm (production time, annual volumes, points of variability) to establish baselines.
- Design and develop reusable AI components (templates, specialised agents, business rule libraries) for high-volume uses.
- Handle end to end a priority internal process: technical feasibility of connections to the information system, prototyping, hardening, and management of error cases.
- Build the evaluation framework: representative test sets, quality criteria (accuracy, completeness, format compliance, confidentiality), automated scoring or human review, and regression detection.
- Measure the gains achieved against the baselines and present them in a usable form to support an investment decision.
- Iterate with pilot users and document the work so that a third party can take it over and evolve it.
Qualifications
- Final-year student of an engineering school or master's programme (Bac+5) in computer science, data or artificial intelligence.
- Python; understanding of language models (how they work and their limitations); prompt engineering; knowledge of RAG and agentic architectures; Git.
- Depending on focus: REST APIs and authentication, or solid foundations in statistics (sampling, comparison of measurements, significance) and knowledge of software testing.
- A first experience on an applied AI project is appreciated.
Additional information
- Tailored career paths (career development, adapted training, mentoring…);
- Having the choice of your mission and personalised support throughout your career;
- Working in an environment where learning is encouraged: certifying training, e-learning, meetUps, coding competitions, career development paths, etc.;
- Being part of expert communities that share their knowledge and experience within our competence centres;
- A friendly environment with many social events (annual party, seminars & team building, lunches and afterworks…);
Your differences are our strengths. That is why Meritis is committed to diversity and non-discrimination. All of our jobs are accessible to people with disabilities. If you have a question or believe you are a victim or witness of discrimination, you can contact [email protected] .