AI Data Scientist

About Arlequin

Arlequin is building the sovereign European alternative to Palantir: an AI platform capable of processing massive data volumes without bias, with high-impact use cases such as fighting disinformation. Founded in late 2024 by Hugo Micheron (researcher in geopolitics and systemic risk) and Antoine Jardin (CNRS research engineer).

The role

You'll join our R&D team of 8, focused on researching, testing, and developing state-of-the-art algorithms that power HUDEX, our platform that lets non-technical analysts interpret narrative dynamics across large-scale digital ecosystems. You'll work alongside experienced researchers, several of whom hold or are completing PhDs, across a range of fields (unsupervised learning, NLP, computer vision, topology, and more).

Stack

• Technical stack:

  • Must have: Python and strong experience in data manipulation / data science / machine learning / development tools and libraries, Git/GitHub

  • Nice to have: LaTeX, MLflow

• Methods & domains:

  • Must have: experience in NLP, working with high-dimensional unstructured data (text, computer vision, speech, multimodality…), clustering and statistics

  • Nice to have: experience with other unsupervised learning methods, deep learning, mathematical knowledge in graph and/or topology

What we're looking for

• A strong scientific mindset and genuine intellectual curiosity
• Able to read scientific literature and produce literature reviews
• A high degree of autonomy
• Versatility and adaptability: you're comfortable researching across many fronts, switching between quick, pragmatic solutions and deeper long-term work, and occasionally stepping outside pure research to support engineering or client-facing needs
• Quick to pick up new algorithms and put them into practice
• Open to junior candidates
• Professional English

Bonus

• Academic publications

• Already comfortable in a startup environment

• Mastery of agentic coding tools and code models

Process (~2 weeks)

1. Screening (15 min)

2. Fit interview (30 min)

3. Technical interview (1h)

4. Founder culture fit (45 min)

5. Reference calls

6. Offer