Data Scientist Permanent PositionNorthern Suburbs, Cape Town   Role Purpose Build novel new applications to enable and assistdata-driven decision making within retail operations using advanced data science techniques. Areasof focus includes but is not limited to behaviour prediction, image recognition and natural languageprocessing. These new applications are typically high profile and have an impact on revenuegeneration and cost reduction. Lead the requirements discovery process in a retail environment Model and frame business scenarios that are meaningful and which impact on critical businessprocesses and/or decisions Work alongside business analysts suggesting other solutions that may beof interest to the business stakeholder. Seek out, explore and manage data relevant to therequirements Leveraging new data collection processes such as IoT (Internet of things) devices orsocial media Collaborates with business subject matter experts to select the relevant sources ofinformation. Develops and executes data strategies that cover: data collection, integration andretention requirements incorporating business requirements and knowledge of best practices. Createmodels using data science techniques, particularly focusing on neural networks  Selectappropriate models applicable to the business problem. Select appropriate infrastructure fortraining (local or cloud-based) and advise on requirements (e.g. CPU, GPU). Train models on prepareddata and monitor convergence and model performance. Optimize and improve model performance by tuninghyperparameters and other techniques. Own the development of a complete application that exploitsthe model  Determine the business process context for the model and determine an applicationboundary that allows a complete solution to be created Works alongside business analyst anddevelopers in defining and create a complete application that leverages models created Focus oncreating complete autonomous systems rather than producing ad-hoc analysis in report format Own theprocess of deploying the data science application and gaining its acceptance by the business Workwith developers and operators to create automated deployment pipeline for models and theapplications leveraging them Defend results that contradict common belief if needed to subjectmatter experts and organisational leadership Provide business metrics for overall project to showimprovements. Contribution to improvements should initially be measured over iterations Own ongoingquality of the solution in terms of data, model and performance Define the validity ofinformation, what information is meaningful and for how long and what other information it relatedto Partners with data steward to define data quality expectations Lead design and deployment offixes to systems needed.  Role Description
Data Scientist Permanent Position Northern Suburbs, Cape Town   Role Purpose Build novel newapplications to enable and assist data-driven decision making within retail operations usingadvanced data science techniques. Areas of focus includes but is not limited to behaviourprediction, image recognition and natural language processing. These new applications are typicallyhigh profile and have an impact on revenue generation and cost reduction. Lead the requirementsdiscovery process in a retail environment Model and frame business scenarios that are meaningfuland which impact on critical business processes and/or decisions Work alongside business analystssuggesting other solutions that may be of interest to the business stakeholder. Seek out, exploreand manage data relevant to the requirements Leveraging new data collection processes such as IoT(Internet of things) devices or social media Collaborates with business subject matter experts toselect the relevant sources of information. Develops and executes data strategies that cover: datacollection, integration and retention requirements incorporating business requirements and knowledgeof best practices. Create models using data science techniques, particularly focusing on neuralnetworks  Select appropriate models applicable to the business problem. Select appropriateinfrastructure for training (local or cloud-based) and advise on requirements (e.g. CPU, GPU). Trainmodels on prepared data and monitor convergence and model performance. Optimize and improve modelperformance by tuning hyperparameters and other techniques. Own the development of a completeapplication that exploits the model  Determine the business process context for the model anddetermine an application boundary that allows a complete solution to be created Works alongsidebusiness analyst and developers in defining and create a complete application that leverages modelscreated Focus on creating complete autonomous systems rather than producing ad-hoc analysis inreport format Own the process of deploying the data science application and gaining its acceptanceby the business Work with developers and operators to create automated deployment pipeline formodels and the applications leveraging them Defend results that contradict common belief if neededto subject matter experts and organisational leadership Provide business metrics for overall projectto show improvements. Contribution to improvements should initially be measured over iterations Ownongoing quality of the solution in terms of data, model and performance Define the validity ofinformation, what information is meaningful and for how long and what other information it relatedto Partners with data steward to define data quality expectations Lead design and deployment offixes to systems needed.  

Skills and Experience
Essential Post graduate science degree in Data Science, Mathematics, with computer science major,or Computer Science, with mathematics major 5 years experience in customer and/or retail analysis 5years experience in Customer and/or retail data science experience with focus on neuralnetworks while building and putting into production applications that leverage these models 2 yearsexperience in at least three of the following:           - Customer behaviour prediction          - Assortment optimisation          - Customer driven marketing          - Supply chain analytics          - Recommendation engines          - Image recognition          - Natural Language          - Digital assistants 2 years - Skill in using Python for data analysis and buildingneural networks 2 Years - Skills in using SQL and scripting tools (RegEx, Awk, Shell scripting) inthe data mining and  cleansing process, particularly for data preparation 2 years - Familiaritywith at least one neural network framework such as Keras or TensorFlow 2 years - Strongunderstanding of data science theories, principles and practices 2 years - Expert knowledge of atleast two of the following:          - Decision trees         - Association          - Transfer learning          - Clustering Email:  Leigh.Duffield@eoh.co.za