* The Data Scientist has full-stack accountabilities across the full value chain of building an industrialized data-science software product:
* Understanding a business problem and its component processes end to end, and identifying opportunities to make decisions more optimally leveraging decision-support tooling
* Efficiently conducting analyses and visualizations to identify valuable opportunities for decision-support and to determine trade-offs between different potential feature implementations
* Prototyping advanced machine learning and optimization models to prove the value of a use case and approach (in Python)
* Delivering features to industrialize machine learning and optimization models in Python using best-practice software principles (e.g., strict typing, classes, testing)
* Build automated, robust data cleaning pipelines that follow software best-practices (in Python)
* Implementing integrations between the core algorithm (machine-learning or optimization) and a workflow orchestration paradigm such as Dagster
* Implementing software in a cloud-based deployment pipeline with Continuous Integration / Continuous Deployment (CI/CD) principles
* Building logging, error handling, and automated tests (e.g., unit tests, regression tests) to ensure the robustness of operationally critical decision-support products
* Deliver features to harden an algorithm against edge cases in the operation and in data
* Conduct analysis to quantify the adoption and value-capture from a decision-support product
* Engage with business stakeholders to collect requirements and get feedback
* Contribute to conversations on feature prioritisation and roadmap, with an understanding of the trade-off between speed vs. long-term value
* Understand and integrate the product into existing business processes, and contribute to the development and adoption of new business processes leveraging a decision-support product
* Communicate feature and modeling approach, trade-offs, and results with the internal team and business stakeholders
* The Data Scientist is also accountable for ways of working fit for an Agile cross-functional development squad, including:
* Using Git-versioning best practices for version control
* Contributing and reviewing pull-requests and product / technical documentation
* Giving input on prioritization, team process improvements, optimizing technology choices
* Working independently and giving predictability on delivery timelines
Skills/capabilities
* Strong knowledge of either machine learning and optimization techniques, incl. supervised (regression, tree methods, etc.), unsupervised (clustering) learning, and operations research (linear, mixed integer programming, heuristics)
* Fluent in Python(required) and other programming languages (preferred)with strong skills in applying DS, ML, and OR packages (scikit-learn, pandas, numpy, gurobietc.) to solve real-life problems and visualise the outcomes (e.g. seaborn)
* Proficient in working with cloud platforms (AWS preferred), code versioning (Git), experiment tracking (e.g. MLflow)
* Experience with cloud-based ML tools (e.g. SageMaker), data and model versioning (e.g. DVC), CI/CD (e.g. GitHub Actions), workflow orchestration (e.g. Airflow/Dagster) and containerised solutions (e.g. Docker, ECS) nice to have
* Experience in code testing (unit, integration, end-to-end tests)
* Strong data engineering skills in SQL and Python
* Proficient in use of Microsoft Office, including advanced Excel and PowerPoint Skills
* Advanced analytical skills, including the ability to apply a range of data science and analytic techniques to quickly generate accurate business insights
* Understanding of the trade-offs of different data science, machine learning, and optimization approaches, and ability to intelligently select which are the best candidates to solve a particular business problem
* Able to structure business and technical problems, identify trade-offs, and propose solutions
* Communication of advanced technical concepts to audiences with varying levels of technical skills
* Managing priorities and timelines to deliver features in a timely manner that meet business requirements
* Collaborative team-working, giving and receiving feedback, and always seeking to improve team processes
Qualifications/experience
* Master’s degree or greater in data science, ML, or operational research, or 2+ years of highly relevant industry experience(required)
* 0-2 years working on production ML or optimization software products at scale (required)
* Experience in developing industrialized software, especially data science or machine learning software products (preferred)
* Experience in relevant business domains (transportation, airlines, operations, network problems) (preferred)
Key interfaces
* Lead Product Data Scientist
* Other Data Scientists
* Business stakeholders and users
* Software engineers (front-end, back-end, DevOps, data engineers)
* Product & change managers
* BA Digital teams (e.g., architects, application support managers)
* External partners and third parties, as required
* ODS Leadership (Head of Data & Analytics, Head of iOps& Optimisation, Director of ODS)
* Key performance indicators
* Model accuracy, performance, and runtime (precision, recall, accuracy)
* Time to develop and deploy features and models
* Data ingestion & processing efficiency and robustness
* Code quality and robustness (e.g., unit test coverage)
* Collaboration and cross-functional teamwork
#J-18808-Ljbffr