Junior AI Researcher

Giotto.ai is a Swiss AI company building advanced intelligence systems for Switzerland and Europe. Our portable AI model and operating system enables organisations to use powerful reasoning capabilities while retaining control over their infrastructure, data and operations.
About the role
We are looking for a talented Master’s student or recent graduate to join us as a Junior AI Researcher.
You will work alongside experienced researchers and engineers to train, evaluate and improve advanced AI models. Your work may involve large language models, reasoning, synthetic data, model evaluation, multimodal learning and task-specific adaptation.
This is a hands-on opportunity to move from research papers and ideas to experiments, prototypes and real AI capabilities.
You will

• Read and discuss recent AI research and help translate promising ideas into experiments.
• Train, fine-tune and evaluate Transformer-based models.
• Contribute to experiments involving reasoning, synthetic data, model adaptation and evaluation.
• Build benchmarks and analyse model performance and failure cases.
• Explore improvements to training data, objectives, prompts and inference strategies.
• Write clean and reproducible research code.
• Document experiments, results and conclusions.
• Collaborate with research, engineering and product colleagues.
• Learn how advanced research is transformed into production-ready AI systems.

You bring

• A completed or nearly completed Master’s degree in computer science, artificial intelligence, machine learning, mathematics, physics or a related quantitative field.
• A solid understanding of machine learning and deep learning fundamentals.
• Practical experience with Python and PyTorch through university projects, a thesis, research work or an internship.
• Familiarity with Transformer architectures and large language models.
• An understanding of probability, statistics, linear algebra and optimisation.
• The ability to read research papers and implement ideas from them.
• Curiosity, analytical thinking and a willingness to learn.
• Clear communication skills and a collaborative mindset.

Nice to have

• Experience with Hugging Face Transformers or Datasets.
• A Master’s thesis related to LLMs, reasoning, multimodal learning, reinforcement learning, interpretability or generative AI.
• Experience with model fine-tuning, LoRA/PEFT, synthetic data or evaluation pipelines.
• Research publications, open-source contributions or relevant personal projects.
• Experience using GPUs, experiment-tracking tools or cloud infrastructure.

You do not need to meet every listed requirement. We value strong fundamentals, curiosity, thoughtful experimentation and evidence that you enjoy solving difficult problems.
Location and work style
This is a full-time position based in Switzerland. We support a hybrid working model, with regular collaboration at our Lausanne office.