Data Scientist Interne

Shaping the future of banking with AI: Distilling and optimizing Large Language Models for AI assistants in finance and banking

The integration of financial services with technology has created an overwhelming amount of structured and unstructured data. To turn this into reliable, real-time intelligence, we need efficient and well-optimized language models, not just raw scale.
Swissquote's Data Science team offers an internship focused on distillation, fine-tuning, and parameter-efficient adaptation (e.g. LoRA) of Large Language Models (LLMs) for production-grade AI assistants.
The main objective is to explore, design, and refine model compression and adaptation techniques to build fast, cost-efficient, and robust AI assistants tailored to the banking and finance domains.

The DS team is integrated in Swissquote's core research group, the Quantitative Research & Solutions (QRS) department, specialized in theoretical and applied research in Quantitative Finance and Data Science.
QRS's mission is to provide best-in-class quantitative and AI solutions by performing research to support every division of the bank in need of an analytical and data-driven approach.

By Joining The Team As Data Scientist Intern You Will :

• Prepare and curate financial text datasets for training, fine-tuning, and evaluation
• Apply and benchmark LLM distillation techniques to build smaller, faster chatbot models
• Use LoRA and other parameter-efficient fine-tuning methods to adapt LLMs to banking-specific tasks
• Design experiments, evaluate model performance, and help integrate optimized models into AI assistant pipelines

• Solid background in data science and NLP
• Prior exposure to, or a strong interest in, LLMs, fine-tuning, and/or model compression
• Coding proficiency in Python (PyTorch / TensorFlow or similar is a plus)
• Interest in the banking and financial sectors
• Creative and curious, comfortable questioning existing approaches
• Organized, self-motivated with excellent communication skills
• Ability to work in a team and communicate with different stakeholders
• Fluent in English