About the job
As a research engineer in the semantic understanding and reasoning group (CR/AIR4) at Bosch Corporate Research, you will develop the next generation of agentic AI systems based on reinforcement learning, with a primary focus on applications in systems engineering. Your work will address how intelligent agents can support and partially automate complex engineering workflows by learning to make structured decisions in environments shaped by constraints, specifications, system models, and long-horizon objectives.
Responsibilities
Design AI agents that interact with engineering artifacts, reason over goals and constraints, and improve behavior through feedback, simulation, and optimization; investigate combining reinforcement learning, hierarchical decision-making, model-based methods, and planning with modern agentic AI architectures to support engineering tasks such as architecture exploration, requirement analysis, system-level trade-off evaluation, validation support, and process optimization; define suitable state and action representations for technical workflows; integrate symbolic and structured knowledge; design reward mechanisms aligned with engineering objectives; build simulation or surrogate environments for safe and efficient agent learning; enable interaction between language-based agents and formal engineering tools to operate across textual, symbolic, and numerical representations; prototype and evaluate methods in realistic use cases with research scientists, AI engineers, and systems engineering experts.
Qualifications
Minimum
excellent MSc in Computer Science, Machine Learning, Robotics, Systems Engineering, Control, or related fields; PhD in Machine Learning, Reinforcement Learning, Agentic AI, Sequential Decision-Making, or related areas; strong publication record in leading AI, machine learning, or autonomous systems venues such as NeurIPS, ICLR, ICML, AAAI, IJCAI, CoRL, RSS, AAMAS, or similar
Preferred
experience with model-based RL, offline RL, hierarchical RL, multi-agent RL, or constrained RL; familiarity with agentic AI architectures that involve goal-directed behavior, memory, tool use, multi-step reasoning, and long-horizon task execution; ability to design agents that learn from interaction, simulation, or structured feedback in complex environments; strong interest in applying AI to systems engineering tasks such as design-space exploration, requirement analysis, architecture optimization, verification support, or engineering workflow automation; familiarity with structured engineering artifacts such as requirements, system models, dependency graphs, simulation outputs, or test specifications; ability to formulate engineering problems as sequential decision-making or optimization tasks; interest in combining formal engineering processes with adaptive AI methods; experience with planning, search, optimization, or decision-making under constraints and uncertainty; familiarity with simulation-based learning and the creation of training environments for agents operating in technical or cyber-physical domains; interest in combining RL with symbolic representations, structured world models, knowledge graphs, or formal methods; understanding of how language-based interfaces and reasoning modules can be integrated into agentic decision systems; solid experience in Python and modern deep learning frameworks such as PyTorch, TensorFlow, or JAX in industrial real-world applications; familiarity with scalable experimentation, distributed training, and evaluation pipelines; experience with Docker, Git, CI/CD, and collaborative software development practices; ability to build reproducible research infrastructure for training, benchmarking, and analyzing agentic AI systems