Senior Data Scientist- Azure ML

Key Skills

Algorithms, AWS, AWS Sagemaker, Azure, Continuous Integration & Continuous Delivery - CI/CD, Data science techniques, Employee life cycle, Git, Google Cloud Platform (GCP), IT, Machine learning techniques, NumPy, Pandas, Python Programming, R Programming, Statistics

Job Description

• Accountable for delivery of data science products in a production scale environment
• Manage the product lifecycle especially relating to data science, and development to support.
• Transform proof of concepts into a larger deployable product in Shell and outside. Rinse and repeat, cross-pollinate ideas across projects and assets
• Partner with P&T Research teams, convert Proof of concepts into scalable data products
• Value estimation and Realization, end-to-end value mapping for deployed data science solutions in production
• Responsible for the supervision and contribution to project work to build machine learning models and digital solutions, providing the technical assurance on end-2-end delivery of data science solutions for Shell assets, operations and new business transformations, in particular, the machine learning component
• Will adhere to the principles of delivering quality by conducting regular health checks, audits, and code reviews to ensure that clean and maintainable production-level code is being produced by project teams
• Will evaluate and benchmark all new asset-related initiatives prior to developing solutions.
• Will keep abreast of new engineering practices, and technologies and continuously improve our Agile practices and build and strengthen relationships and alliances in academia, IDT and industry
• DIY Software & Tooling tasks including Python, AzureML, AWS Sagemaker
• Ensure that the design is fit for purpose for DIY Data Science development needs while meeting controls to safeguard secure and reliable operations
• Create/embed learning curriculum/certification approach for DIY Data Science development
• Engage the community to embed the use of the DIY Data Science toolset, including evangelizing expected ways of working and use of tooling's such as source control, CI/CD and static analysis to meet software engineering policies in the IT Control Framework
• Work with CDOs to agree on data access policies for DIY Data Science usage
• Operationalize DIY Data Science ways of working across LOD1 and LOD2 organizations.
• Create a broader tooling strategy for DIY Data Science in coordination with the Software Engineering portfolio manager and segment architect
• Work with Software Engineering & Architecture leads to developing and delivering against the Tooling Epic of the product backlog, as well as coordinating with any technology-related delivery in other Epics (eg Zoning & Controls, Service).

What we need from you

• A 4-year graduate or master’s degree in IT or MSc or equivalent in Statistics, Mathematics, Econometrics or similar discipline with at least 7 years experience on data science projects.
• Experience in developing, deploying, and supporting models in production
• Experience in developing end-to-end models using machine learning for predictive modelling in a business/industry environment, i.e. feature engineering, model creation and evaluation
• Ability to write clean, elegant and maintainable production-level code in Python and R, and experience in version control, testing, refactoring
• Extensive experience working in small, empowered, delivery-focused, cross-functional and collaborating teams and working in a virtual team and with virtual stakeholders, explaining technical concepts
• Experienced coder in data science languages such as R, Python (Pandas, Scikit Learn, etc.) and Matlab and working in big cloud environments (e.g. AWS, Azure, GCP, etc.)
• Has a firm grasp and demonstrable experience on the principles of the model development lifecycle and producing machine learning models that stand up to statistical rigour
• Has good knowledge and understanding of Shell.ai, GIT, Azure Machine Learning, AWS SageMaker, CI/CD pipelines
• Have experience in at least one out of the following three: Probabilistic Modelling (e.g. General Linear Models, Stochastic processes, Bayesian Networks, Hidden Markov Models, Hierarchical Modelling, Time Series Modelling, etc.), Algorithms and optimization (e.g. linear programming, multi-objective optimization, meta-heuristics, etc.)
• Knowledge of Bayesian Statistics and Bayesian/Probabilistic modelling