Master Thesis: Verifiable Reward RL (RLVR) for Multilingual Agentic Tool & Function Calling
AI Sweden
Summary
For LLMs to serve effectively in public administration and enterprise workflows, they must accurately invoke external APIs and structured tools (Function Calling). While SFT provides basic tool-use syntax, Reinforcement Learning with Verifiable Rewards (RLVR) is required to optimize multi-step agentic trajectories and error recovery.
Job description
Introduction
For LLMs to serve effectively in public administration and enterprise workflows, they must accurately invoke external APIs and structured tools (Function Calling). While SFT provides basic tool-use syntax, Reinforcement Learning with Verifiable Rewards (RLVR) is required to optimize multi-step agentic trajectories and error recovery.
Project Background and Problem Statement
OpenEuroLLM is developing the oellm-rlvr framework to support agentic and function-calling rollouts.
- This thesis asks: Does RLVR training on tool-execution environments (e.g., Tau-Bench, BFCL) significantly improve multi-step function-calling accuracy and schema adherence in Swedish and English compared to SFT alone?
Outline
- Literature Study: Review agentic environment design, function-calling benchmarks (BFCL, Terminal-Bench), and RLVR mechanics.
- Implementation: Adapt oellm-rlvr to execute tool-use rollouts with deterministic execution verifiers on Prelude 9B.
- Evaluation: Benchmark tool-calling precision, schema validity, and Model FLOPs Utilization (MFU) across single-turn and multi-turn administrative workflow.
Who we're looking for
We are seeking curious, self-driven MSc students eager to work at the frontier of open-weight European AI research (LLMs). You thrive on empirical discovery, design rigorous experiments, and let data challenge your assumptions.
- Ongoing Master's studies in Computer Science, Data Science, Machine Learning, Engineering Physics, or a related quantitative field.
- Proficiency in Python and hands-on experience with modern deep learning frameworks (PyTorch, Hugging Face ecosystem).
- Familiarity with LLM post-training alignment (e.g., SFT, DPO, RLHF/RLVR) or context-extension, alongside comfort running distributed GPU training in Linux/HPC environments.
Requirements
- Ongoing Master's studies in Computer Science, Data Science, Machine Learning, Engineering Physics, or a related quantitative field
- Proficiency in Python
- Hands-on experience with modern deep learning frameworks (PyTorch, Hugging Face ecosystem)
- Familiarity with LLM post-training alignment (e.g., SFT, DPO, RLHF/RLVR) or context-extension
- Comfort running distributed GPU training in Linux/HPC environments
Company overview
AI Sweden is Sweden's national center for applied artificial intelligence, established in 2019 with a mission to accelerate the adoption and implementation of AI technologies for societal benefit, economic competitiveness, and national advancement. As a collaborative research institute, it brings together over 170 partners from diverse sectors including government agencies, universities, research institutions, startups, and global corporations. The organization operates as a catalyst for AI innovation, focusing on practical applications across healthcare, energy systems, municipalities, and various industries. With a core team of over 80 professionals comprising AI experts, data scientists, research engineers, linguists, mathematicians, policy experts, and change leaders, AI Sweden facilitates knowledge sharing, talent development, and infrastructure building. The institute operates on a not-for-profit basis, funded by Sweden's innovation agency Vinnova, the Swedish Agency for Economic and Regional Growth, European Regional Development Fund, Västra Götalandsregionen, and partner contributions. Through its AI Labs, talent programs, workshops, and collaborative projects, AI Sweden addresses critical societal challenges while bridging the gap between cutting-edge research and real-world implementation, positioning Sweden as a leader in responsible and impactful AI development.