Lead Data Scientist, Wise

• Experience implementing, training, testing and evaluating performance of Machine Learning systems
• Strong Python knowledge. A big plus for proven familiarity and experience with OOP principles
• Knowledge and experience developing Unsupervised Learning methods
• Experience with statistical analysis, and ability to produce well-designed experiments
• A strong product mindset with the ability to work independently in a cross-functional and cross-team environment
• Good communication skills and ability to get the point across to non-technical individuals
• Strong problem solving skills with the ability to help refine problem statements and figure out how to solve them
• We want to work with the best people, no matter their background. So if you’re passionate about learning new things and keen to join our mission, you’ll fit right in

Desirable

• This role is a unique opportunity to work on the intelligence system at the core of our contact automation system
• Your work will make our system aware of the different problems our customers encounter and eventually develop the ability to help customers resolve these issues
• What you build will have a direct impact on Wise’s mission and millions of our customers
• In the Support squad we are aiming to create a system that can power an automated “Wise Assistant” system that can answer most customer questions within the chat interface effectively, as well as support our agents in answering more complex questions
• To achieve our targets we need to have applied this system effectively at scale, across the large majority of our contacts working seamlessly within the chat interface
• In the near-term our Data Science members are primarily focused on developing the foundation of the customer problem understanding system. There are many interesting research angles, especially within the NLP domain
• Analysing contact data to identify patterns and uncover underlying structures
• Creating automated algorithms for extracting information from real customer interactions
• Developing the system’s ability to learn which data points are key to resolving specific customer problems
• Innovating prompt development to optimise the performance of LLM-based parts of the system and background processes
• Developing models to map customer contacts to defined structures
• Evaluating our customer problem understanding system against internal and external benchmarks
• Identifying and categorising system errors, and suggesting technical solutions to rectify these errors
• Fine-tuning system settings to achieve an optimal balance between precision and recall
• Providing data-driven insights on potential outcomes under various scenarios
• Collaborating with operational teams to refine processes, ensuring effective feedback integration into our automation systems
• Designing and managing projects that utilise excess operational capacity, such as manual data labelling for model improvement
• Integrating active learning strategies to continuously improve model accuracy through feedback loops
• Packaging algorithms into deployable libraries/objects and transitioning them from staging to production environments
• Implementing and maintaining scheduled processes for data gathering and model retraining using automated pipelines
• Maintaining production-grade Python services