Data Science Intern – Lightsource bp
As a Data Science Intern at Lightsource bp, you will work on high‑impact optimisation and decision‑science problems that underpin our energy storage and grid‑integration strategy. You’ll contribute to research, design, experiment, implement and test approaches (modelling, simulation, reinforcement learning and optimisation) that support improved operational decision‑making for battery energy storage systems (BESS).
This internship is designed for a PhD‑level researcher who wants to apply cutting‑edge methods to real‑world energy systems within a global renewable energy organisation.
Details
Duration: 1 June – 21 August 2026 (12 weeks)
Location: London or Milan (Hybrid Working – 3 days per week in office)
Type: Full‑time internship
Education Required: Currently pursuing a PhD (or final‑year PhD candidate) in a quantitative field; Master’s degree required as minimum. Alternative educational paths are considered if equivalent experience is demonstrated.
Responsibilities
* Develop and evaluate reinforcement learning (RL), (constrained) optimisation, or simulation models to support BESS dispatch strategy.
* Prototype algorithms for sequential decision‑making under uncertainty, incorporating market, operational, and physical system constraints.
* Analyse large time‑series and operational datasets to extract actionable insights that support energy trading, grid services, and storage optimisation.
* Collaborate closely with data scientists and product teams to design experiments and validate models.
* Communicate findings clearly to both technical and non‑technical stakeholders through documentation, visualisations, and presentations.
* Contribute to reproducible, high‑quality research code and experimentation pipelines.
* Stay current with emerging methods in RL, optimisation, and energy systems research.
Experience & Knowledge Required
* Experience conducting research or applied work in one or more of: reinforcement learning, control systems, (constrained) optimisation, probabilistic modelling or simulation.
* Strong understanding of sequential decision‑making frameworks (e.g., MDPs, policy optimisation, planning).
* Experience working with large datasets, experimental design, and empirical model evaluation.
* Understanding of numerical methods, statistical learning, or stochastic processes.
* Familiarity with energy systems, grid flexibility markets, or storage optimisation is a plus but not mandatory.
Skills Required
* Strong proficiency in Python and its scientific computing toolkit (JAX, Pyomo, NumPy, Scipy, Statsmodels, Pandas/Polars and similar).
* Solid knowledge of PyTorch, Gymnasium and one or more of the following reinforcement learning frameworks: StableBasleine3, TorchRL, RLlib, SKRL.
* Ability to design and evaluate RL agents or optimisation algorithms.
* Strong analytical and mathematical reasoning skills.
* Ability to communicate complex technical findings clearly and concisely.
* Experience with scientific coding best practices (version control, experimentation tracking, reproducibility).
* Ability to work in a research environment while following agile ways of working (continuous iterative delivery and improvement).
* Curiosity, autonomy and good organisational and time‑management skills.
Education Required
* Currently pursuing a PhD (or in the final years of a PhD) in Machine Learning, Computer Science, Applied Mathematics, Engineering, Operations Research, or a related quantitative field.
* Other educational paths are considered if the candidate can demonstrate extensive experience in one or more of the above areas.
Benefits
* Competitive salary with annual bonus and retention bank.
* Health insurance, pension and other local benefits.
#J-18808-Ljbffr