Data Scientist Internship

Your mission Our Engineering team Our Engineering team is responsible for developing our SaaS platform and building a comprehensive and user-friendly product. Pigment engineers participate in the entire application development lifecycle, focusing on design, coding, and keeping the production platform up and running. They can be specialized, but there is no strict separation between infrastructure, backend, and frontend. We value user-centricity and pragmatism: we choose the most relevant tools for the problem we have to solve, understanding the strengths and constraints of each technology. Our engineering culture also values curiosity, humility, trust, ownership, and team spirit. 👉 Curious to see what we’re building? Check out our Tech Blog! As a Data Scientist Intern, you will help us push the boundaries of what AI can do inside Pigment, focusing on how our AI Modeler builds and maintains planning models. You’ll work closely with our AI engineering and data science teams to design, test, and evaluate LLM-based systems that turn loosely structured inputs into reliable, working Pigment models and formulas. Project Overview: Making Pigment’s AI Modeler Smarter Depending on your profile and interests, you will focus on one of these topics: 1. Formula Writing from Specifications Have the Modeler write a specification of a formula (what it should do, and the tests it must pass), then build a small Formula Writer model that produces a correct formula from it. 2. Building Models from External Documents Turn amorphous inputs (long PDFs, call transcripts, existing applications, templates) into structured Pigment applications: write the tests first, then iterate on a DSL representation until they pass. 3. Building Models from Databases Given an existing database, determine which blocks and transformations to create to make sense of the data and integrate it as a Pigment model. As an intern, you will: • 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. Technical Stack Our current AI and engineering stack includes: • 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. Work Environment • Hybrid: possibility to work partially remotely • Office in the heart of Paris (Paris 9) • Choose your working environment (hardware / OS) • A competitive remuneration package • Opportunity to continue your journey with us after your internship Technical stack • Our current AI and engineering stack includes: • 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.