The Desk Data Scientist at Vitol has impact across the full workflow of data science: from working with business stakeholders to help define the project, to data collation and processing, exploratory analysis, model selection and tuning, and implementation of production models or custom visualizations. We extract value from data in many different ways, and experience exploring and extracting insights from data is just as important as the ability to build complex models. When driving business outcomes with data, a good visualization is often just as important as a small MAE. The data we work with is often complex and messy, and a successful Desk Data Scientist at Vitol is always willing to roll up their sleeves and delve into the data in order to achieve results.
The Desk Data Scientist will regularly engage directly with traders, brokers, analysts, and researchers across nearly all aspects of Vitol in order to understand the data that goes into internal models and the business problems that need to be solved, but the day-to-day collaboration will also be with the small and highly agile Data Science team.
Responsibilities include:
- Being an energetic and enthusiastic member of a team uniquely positioned to bring the power of data to bear on the inner workings of the energy industry
- Ability to relate effortlessly with the commercial side of the business, continuously looking for new ways how AI/DS can create value for Vitol
- Exploring new datasets, extracting insights, and visualizing the results
- Standardizing and structuring data in clear, extensible, easily understandable ways
- Applying machine learning models in novel ways to heterogeneous data sets to help optimize Vitol’s diverse businesses
- Helping review the code and experimental design of other Data Scientists to ensure a high level of production quality
- Building data expertise with domain experts in specific aspects of Vitol’s operations
- Aiding in the production-level implementation of the above models and custom visualizations to help drive business value