Praktikum im Bereich AI & Deep Reinforcement Learning in der Intralogistik d/m/w

Airbus
Germany
Workplace: OnsiteInternshipFunction: Education & TrainingSkills: ["Independent","Analytical problem-solving","Solution-oriented","Debugging"]

Support the developer team in optimizing and validating a simulation environment for autonomous control of production cranes. Adapt and extend an existing Python codebase using PyTorch, transitioning a DDQN to a dynamic, graph-based architecture with heterogeneous GNNs. Build and test action-masking mechanisms to prevent collisions on a single-track rail infrastructure, run training iterations, refine reward shaping, and compare DRL-GNN performance to heuristics and exact solvers while visualizing results.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Airbus
Airbus
4 days ago

Praktikum im Bereich AI & Deep Reinforcement Learning in der Intralogistik d/m/w

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 5 hours agoStatus: Live

Job Summary

Support the developer team in optimizing and validating a simulation environment for autonomous control of production cranes. Adapt and extend an existing Python codebase using PyTorch, transitioning a DDQN to a dynamic, graph-based architecture with heterogeneous GNNs. Build and test action-masking mechanisms to prevent collisions on a single-track rail infrastructure, run training iterations, refine reward shaping, and compare DRL-GNN performance to heuristics and exact solvers while visualizing results.
Location: Germany
Workplace: Onsite
Employment Type: Internship
Job Function: Education & Training
Seniority: Intern level

Key Responsibilities

  • •Adapt, extend, and test an existing Python codebase using PyTorch.
  • •Transform a flat DDQN state representation into a dynamic, graph-based architecture using heterogeneous GNNs.
  • •Model cranes and mounting stations as graph nodes and spatial dependencies as graph edges.
  • •Develop and test action-masking mechanisms to avoid collisions and invalid actions on single-track rail infrastructure.
  • •Run training experiments, refine reward shaping, analyze model convergence under varying order loads, and compare against heuristics and exact solvers; visualize results and training statistics.

Key Requirements

  • •You are a currently enrolled student (m/w/d) in Informatics, Data Science, Mathematics, Automation Engineering, or a comparable program.
  • •Strong knowledge of Python and practical experience with PyTorch.
  • •Basic prior knowledge in Deep Reinforcement Learning, Graph Neural Networks (e.g., PyTorch Geometric), or mathematical optimization is advantageous.
  • •You work independently with a solution-oriented approach and analytical skills for debugging code.
  • •Required internship/practice semester according to your degree program guidelines.
Experience:AIMachine learningDeep reinforcement learningGraph neural networks
Education:
Skills:IndependentAnalytical problem-solvingSolution-orientedDebugging
Tech Stack:PythonPyTorchDeep Reinforcement LearningGraph Neural NetworksPyTorch GeometricDDQNHeterogeneous GNNsAction maskingReward-ShapingGoogle OR-ToolsCP-SATFIFOSimulation

Company Brief

Airbus
Designs, manufactures, and sells commercial aircraft, helicopters, defense and space systems, and related services worldwide. Airbus is a leading aerospace and defense company delivering integrated solutions for civil and military aviation customers.
Industry: Aerospace Manufacturing
Company Size: Enterprise (1,001+ employees)
Revenue: USD 1B+
Growth: Public Company
Valuation: Public Company (Market Cap in USD)
Funding: IPO / Publicly Listed
Headquarters: Toulouse, France
Founded: 1970
WebsiteLinkedIn