You design and evaluate LLM-based systems in Paris that help Pigment's AI Modeler build reliable planning models and formulas.
ApplySign-up requiredJob description
You join Pigment's Engineering team, which builds its SaaS planning platform. You work with the AI engineering and data science teams on the AI Modeler, the tool that builds and maintains planning models.
Depending on your profile, you take one of three topics: writing formulas from specifications, building models from external documents, or building models from databases. You learn to define tests, compare approaches and document trade-offs between quality, cost and latency.
The office is in Paris 9, with partial remote work possible. The internship can lead to continuing with Pigment afterwards.
What you'll do
- Define success criteria, test suites, and evaluation metrics for your topic.
- Design and iterate on LLM-based approaches (prompting, agentic workflows, intermediate representations).
- Benchmark quality, reliability, and cost/latency, and document trade-offs.
- Work with engineers and data scientists to build evaluation pipelines and share findings.
Who they're looking for
- Languages & Frameworks: Python, LangGraph
- Data & ML Infrastructure: Weights & Biases, Hugging Face Hub, Google Cloud Platform (GCP), Docker, Kubernetes
- Databases: PostgreSQL
- CI/CD & Experimentation: CircleCI, MLflow, and internal orchestration tools
- We don’t expect you to know them all. What matters most is your ability to learn quickly, experiment rigorously, and translate data insights into actionable results.
Prepare your interview
- Learn what Pigment does, the sector it works in and the projects on its website.
- Go through each task listed here and work out why the team needs it done.
- Prepare a real example for each task, then tie it back to what the company does.
Written by Edu Swap from the listing and, where there is one, the company website. Check the listing to make sure it's still open.