Python Job: Data Scientist, Digital Ethics and Responsible AI

PURPOSE OF JOB

To build the building of explainable and interpretable artificial intelligence models, to glassbox predictive models, test them and deploy them at scale to make predictive models more robust, fair and human centric in fast changing digital world where algorithms are quickly becoming our new digital co-workers.

OUR TEAM WITHIN IKEA

We set the digital ethics policy and group rule at IKEA leading in an uncertain and fast moving environment. We drive the development, provision and operation of Algorithms to ensure robustness, fairness and human centricity in customer and co-worker facing digital products to drive operational excellence and mitigate reputational , business and legal risk.

At the same time, we provide a home for digital skills to both develop and extend the technical, people and business skills needed to ensure continuous development and growth of our digital capability.

WHAT YOU’LL BE DOING DAY TO DAY

• Responsible for exploring , examining and deploying meta Algorithms for extracting why algorithm works, where does it not work, and what can make the algorithms change their behaviour with objective of making algorithms more transparent, robust and rational, in line with Digital ethics policy.

• Influence modeling decisions by comparing explainability in inference pipelines through extensive knowledge of the credit assignment in multi modal models including text, vision and clickstream data

• Develop and deploy different tools to extract explainability, interpretability, auditability and steerability from simple to complex models

• Develop and deploy active learning strategies by explaining which input data points contribute more to prediction accuracy

• Develop algorithms to generate rule based synthetic data to benchmark predictive models’ predictions

• Develop metrics to calculate data drift at inference time with established correlation with performance

• Work closely together with other data scientists and cross functions to produce all required design specifications and ensure that algorithmic solutions work together to mitigate risks related to algorithm behavior

• Work across initiatives within INGKA Group, steering towards explainability -driven algorithms

KNOWLEDGE

• Deep Knowledge of optimization techniques and their pros and cons, deployment patterns, as well as visualization tools e.g. seaborn, matplotlib

• Extensive knowledge of Neural networks, Bagging and Boosting methods

• Extensive Knowledge of machine learning development frameworks (Python,Pytorch, SHAP)

• Experience with MLOps and agile development practices, in which requirements and solutions evolve through the collaborative effort of self- organizing cross-functional teams

• Experience in development of industrialised analytical software products

• Knowledge of IKEAs corporate identity, core values and vision of creating a better everyday life for the many people

JOB – SPECIFIC CAPABILITIES

• Performs work with care, accuracy, and attention to detail and strives to meet quality standards

• Ability to work with large amounts of raw data and confident in working with different programming languages

• High intellectual curiosity with ability to acquire new knowledge and skills and champions new concepts, methods, digital systems and processes to improve performance

• Review user-story prioritisation as appropriate based on new ideas, approaches, and strategies

• Collaborates with others (incl. virtual) to achieve goals

• Ability to understand the complexity of IKEA business and the role of data and information as an integrated part of the business

MOTIVATION

• Passionate about the possibilities of artificial intelligence and its value for business

• Deeply motivated to continuously develop and enhance programming skills and knowledge of machine learning and AI applications

• Energized by working both independently and interdependently with the architecture network and cross functions

• Appreciate the mix of strategic thinking and turning architecture trends into practice

• Share and live the IKEA culture and values

LEADERSHIP CAPABILITIES

• Self driven to take initiatives

• Ability to simplify abstract and complex ideas

• Ability to communicate technical topics to non technical audience

KEY PREVIOUS EXPERIENCES AND PROVEN SKILLS

• M.Sc. in (computer) engineering, computer science, econometrics, mathematics or equivalent Minimum of 2 years of experience as a data scientist in domains requiring explainability such as medicine, banking or insurance.

• Extensive knowledge in large datasets and distributed computing architectures

• Extensive knowledge in coding in Python and development of artificial intelligence applications Experience in statistical analysis and statistical software

• Experience in data processing and –analysis

• Demonstrable experience in working in an Agile or DevOps working set-up

• Preferred: Experience in software engineering

• Preferred: Knowledge of IKEA Brand, culture, values and way of working