Job Title: Data Scientist Location: London, UK (Remote) Job Type: Full-time contract, 6 month Responsibilities: • You will collaborate closely with a team comprising data scientists, data engineers, product developers, and analytics-focused consultants. • Your main focus will be the optimization and automation of a LLM based text classifier for transformation initiatives. The set-up today is ~70% accurate and your main focus will be to improve accuracy through both simple methods such as prompt changes as well as advanced methods (RAG search shortlisting & re-ranking, text quality filters, LLM parameter tuning, multi agent approaches). • Next to this you might also work on broader topics such as: descriptive analytics, predictive models (e.g., boosted trees), and build products that adhere to MLOps best practices, ensuring they are both maintainable and deployable. By doing so you will help bring advanced analytics capabilities into one of Client’s flagship products. Your work will be the backbone for how Client runs future Transformations, leveraging data science assets, to improve the odds of success for our clients. In this role, you will be mainly responsible for the following: • Prompt engineering: you will work and iterate on new prompts to get the best out of the LLM classifying system. • LLM optimization and evaluation: Fine-tune and evaluating the performance and efficiency of Large Language Models, leveraging the latest advancements in neural network architectures and machine learning techniques • Design & build creative approaches to further optimize accuracy & reduce cost: Next to optimizing the LLM through prompts and settings you will design & test other approaches such as pre-filters, non LLM based text models in parallel, multi agent set-ups etc… • Build for scale: Together with the tech lead(s) you will define how to maintain and scale our LLM classifier pipeline. The classifier pipeline will, in the long term, run as a on-demand batch process, and will, once stable, be refactored with scalability in mind. • Coach & help other colleagues: Coach & help you peers when needed, we deliver as a team. • Deploying pipelines at scale: you will be exposed to and later lead efforts to deploy pipelines & LLM processing in cloud environments, where needed with the support of more specialized team members. Qualifications: • Bachelor's degree in the field of computer science, machine learning, applied statistics, mathematics or related field required. Advanced degree is a strong plus • At least one prior experience with LLMs, a demonstrable good understanding of key levers to optimize them and a keen understanding of the latest developments. • Strong data translation and presentation skills, able to clearly and effectively communicate complex analytical and technical content will be required in this role. • Experience data science, statistical modelling (e.g. advanced regressions, clustering, classification models) and machine learning techniques (e.g. random forest, support vector machines, gradient boosting, XGBoost) is a very strong asset. • Familiarity with neural network architectures (Transformers) is nice to have, deep learning libraries (TensorFlow/PyTorch) and machine learning libraries (scikit-learn) • Programming experience in the following languages: Python and SQL • Experience in leading complex engagements to deploy advanced analytics and data science methods at scale in real world organizations is very valued. • Exposure to tools like Excel, VBA, Snowflake and Tableau is a plus