Data and AI Internship

Internship - Paris - Full time - 4 to 6 months

Omaha Insights is bringing AI & ML in their production pipeline. In this sector data quality is crucial, increasing the need for control and rigor when implementing new solutions. This role is very hands-on and sits at the intersection of data & AI engineering, finance and software engineering. 

You’ll work within the R&D team, where financial and data related research happens. This team deals with data production and control, turning raw accounting figures into valuable insights. 

Here at Omaha we like to work at a fast pace and bring ideas to life quickly. This project is highly exploratory and requires creativity, curiosity and a detail-oriented mind. 

What you'll do

• Bring LLMs into production to better understand and classify the data (Agentic AI, LLM-as-judge, RAG, …),
• Normalize data from multiple accounting standards (30+ countries),
• Build Golden Cases, benchmarks and evaluation methods for quality control and error catching,
• Bring ML algorithms to life (mostly classification), deal with outliers,
• Create features directly with the Product/Dev teams to improve the UX of our platform,
• Document your rationale and edge cases clearly enough that decisions can be reviewed and discussed.

What we look for

We're prioritizing technical depth and independent judgment over a specific pedigree. Because this role has real latitude to shape our data methods, we are looking for someone who can propose a direction, not just execute a spec:

• Strong Python fundamentals (Object-oriented programming, clean-coding principles), comfortable owning a module rather than a script,
• Genuine ML/data science depth (classification, embeddings, evaluation methodology), you understand why a technique works, not just how to call it,
• Evaluation rigor: you instinctively want to know if a change actually improved precision/recall/F1 before you trust it, and you build the benchmark to check rather than eyeballing it,
• Comfortable proposing technical direction: this is exploratory work, we want someone who pushes back with "have you considered X" rather than waiting to be told what to build,
• Strong communication skills: you'll be expected to explain your reasoning and flag uncertainty proactively, not wait to be asked,
• Nice to have: exposure to LLMs, agentic workflows or RAG. We’re expecting curiosity and willingness to learn and experiment.
• Nice to have: prior finance/accounting experience.

How we'll work

• A well-scoped project with a clear, self-checkable ground truth, so you can build independence early,
• Your tutor and the team are available for check-ins as needed, don't hesitate to reach out rather than sit on a blocker,
• Clear escalation guidelines from day one, so you know exactly when to flag something rather than guess.