Master Thesis in Continual Learning with Agentic Memories

Master Thesis in Continual Learning with Agentic Memories

Company: Bosch Group Location: Germany Contract: Internship Duration: 6 months Work Model: Hybrid (70% remote, 30% in-office)

About the Role

Are you a motivated Master's student looking to explore the cutting edge of machine learning? Bosch is offering a unique thesis opportunity focused on continual learning within agentic systems. You will investigate how machine learning systems can evolve and maintain knowledge over time, addressing challenges like growing context sizes and high training costs. This role offers the chance to work with common Large Language Models (LLMs) in practice-oriented settings, improving agent working memory and token efficiency.

Key Responsibilities

• Conduct a comprehensive literature review on agent memory, analyzing existing benchmarks, datasets, and methods, with a focus on continual learning.
• Adapt existing benchmarks or implement new ones for Bosch-related use cases.
• Write code to apply LLMs in an agentic setting, with a specific emphasis on agent memory.
• Derive and implement methods to enhance the memory of continually learning agentic systems.
• Rigorously evaluate the performance of developed approaches on academic benchmarks and Bosch use cases.
• Analyze the scalability, robustness, and deployment potential of your solutions.
• Work within a tight project timeline, with the encouragement to submit findings to major machine learning conferences.

Requirements

• Currently pursuing a Master's degree in Computer Science, Mathematics, Machine Learning, or a related field, with a strong academic record and a focus on machine learning.
• Strong academic background in machine learning and applied mathematics.
• Solid programming skills in deep learning using PyTorch.
• Proficiency with Git for version control.
• Familiarity with job scheduling systems.
• Practical knowledge of agentic systems and their implementation in a research context.
• Experience working with LLMs using PyTorch and Python.
• A proactive, independent, and research-oriented approach to problem-solving.
• Ability to work effectively under deadline pressure.
• Keen interest in independent problem-solving.
• Fluent in English; beginner level in German is a plus.
• Must be enrolled at a university for the duration of the thesis.

What We Offer

• A challenging and impactful thesis project at a leading technology company.
• The opportunity to contribute to state-of-the-art research in machine learning.
• A hybrid work model offering flexibility.
• The chance to have your research potentially published in major conferences.
• A collaborative and inspiring work environment.

Start Date: To be agreed upon. Application Documents: Please include your CV, transcript of records, examination regulations, a list of previous code projects with brief descriptions, and, if applicable, a valid work and residence permit.

Bosch is committed to diversity and inclusion, welcoming all applications regardless of gender, age, disability, religion, ethnic origin, or sexual identity.