Thesis project

We are Zeekr Technology Europe. Powered by engineers and creative minds from all over the world. Located in Gothenburg, Sweden, one of the leading mobility innovation grounds worldwide. Established in 2013 as CEVT and began with the mission to create a revolutionary modular vehicle architecture, CMA. As of 2024, CEVT is now Zeekr Technology Europe, signaling our belonging within Zeekr. Delivering intelligent, sustainable, and user centric technology and software for tomorrow's electric mobility today.
What is GEPA?
GEPA (Genetic-Pareto) is a framework for optimizing arbitrary systems composed of text components like: AI prompts, code snippets, or textual specs against any evaluation metric. It employs LLMs to reflect on system behavior, using feedback from execution and evaluation traces to drive targeted improvements. Through iterative mutation, reflection, and Pareto-aware candidate selection, GEPA evolves robust, high-performing variants with minimal evaluations, co-evolving multiple components in modular systems for domain-specific gains.

Project Description
This project investigates methods to systematically improve how large language model (LLM) agents use external tools over multiple conversational turns. The work will design and evaluate an automated prompt-optimization pipeline that iteratively refines agent instructions, tool descriptions, planning guidance, error-handling strategies, and termination criteria.

The student(s) will define representative multi-turn tasks, collect execution traces, derive feedback signals, and compare optimized prompts against strong hand-crafted baselines. Expected outcomes include: (1) a reusable framework for prompt optimization in tool-using agents, (2) measurable improvements in task success and reliability, and (3) insights into which prompt components most strongly influence performance.

Key Focus Areas

• Multi-turn agent design and evaluation (planning, tool selection, argument formatting, recovery, stopping).
• Automated prompt optimization (generation of candidate prompts, feedback-driven refinement, acceptance criteria).
• Benchmark setup and metrics (task success rate, accuracy/quality scores, efficiency, robustness across tasks).
• Trace collection and analysis (capturing tool calls, inputs/outputs, errors, and decisions to inform revisions).
• Ablation studies to identify high-impact prompt components and generalization across domains.

Your Skills And Background

• Ongoing MSc in Computer Engineering or Information Engineering.
• Strong Python programming and experimental rigor.
• Familiarity with large language models, prompt engineering, basic ML evaluation practices and analyze multi-turn agent behavior.
• Ability to design datasets/benchmarks, work with JSON logs.
• Meritorious: Experience with retrieval-augmented generation, model context protocol, semantic kernel, agent orchestration

Why you should join Zeekr Tech Eu
We are engineers, developers, and innovators from around the world. Joined together by entrepreneurship, our unique blend of global culture, and a belief in a smarter more sustainable future. At Zeekr Tech Eu we fast-track innovation and transform ideas into pioneering technology solutions, doing your master thesis here is no different. We are convinced that a thesis project is a major contribution to our innovation capabilities and long-term development. You'll have a great opportunity to use your skills and creativity to push the boundaries of what´s possible.

What Happens When You Apply
If this sounds interesting and you match the requirements, please don't hesitate to submit your application with a CV and cover letter. Shortlisted candidates will be contacted for an interview to further discuss the project's details and expectations.

This thesis project is intended for 2 students
Supervisor:
Mikel Broström, Tech Lead AI, , please contact for more information about the project

Starting date:
January 2026

Last application date:

Apply today. We will perform ongoing selection during the application period. We look forward to hearing from you

Please note that due to GDPR regulations,
we can only accept applications sent through the recruitment system
, not via email or other channels.