Research scientist intern

What Makes You a Good Fit Bachelor’s or Master’s degree in mathematics, statistics, physics, computer science, or another highly quantitative field Strong knowledge of probability and statistics (e.g., machine learning, time-series analysis, pattern recognition) Prior experience working in a data-driven research environment Experience with translating mathematical models and algorithms into code (Python) Independent research experience: you can own a problem end-to-end Ability to manage multiple tasks and thrive in a fast-paced team environment Excellent analytical skills, with strong attention to detail VoltaBack is a B2B2C SaaS startup reinventing the experience of driving a company car. Our first product tackles a real headache, home charging reimbursement, by combining EV data with smart-meter energy data, no hardware required. And this is just the beginning. Over this decade, 35 million company cars in Europe will become both electric and connected. Together these shifts change what a fleet is: it will run on data instead of fuel cards, consume European electricity instead of imported fuel, and eventually earn money rather than spend it, by bringing thousands of batteries into electricity markets. Our ambition is to be the operating platform companies run these connected fleets on. Proprietary algorithms sit at the core of the product: we collect millions of real-world data points every day from vehicles and smart meters, and turn them into solutions to existing problems, insights, and decisions our customers can act on. We're hiring a research scientist intern to improve the performance of our ML and physics-based algorithms. Three algorithm families are currently in production: home charging, trip detection and predictive maintenance. As a research scientist intern, you will work on one of these topics. The goal of this internship is to improve the accuracy of our algorithms against several benchmarks. This is a research-focused role centered on ML and computational modeling. You will: Explore new formulations : challenge the existing physical and statistical models, propose alternative approaches (non-linear regression, deep learning on time series, etc.) Test new algorithms : design and run rigorous experiments, measure impact against ground truth using our evaluation framework Work on the features : engineer better signal representations from raw time series Communication is key. You will report directly to the CTO and present results and project progress regularly. The goal is to maintain a structured documentation practice while exploring as many ideas as possible. Keep in mind : your algorithms will then be put into production by the software engineering team. Clean code, clear documentation, and reproducible experiments are essential. Hybrid workspace Swile for lunch