AI-ML Researcher

About AkzoNobel

Since 1792, we’ve been supplying the innovative paints and coatings that help to color people’s lives and protect what matters most. Our world class portfolio of brands – including Dulux, International, Sikkens and Interpon – is trusted by customers around the globe. We’re active in more than 150 countries and use our expertise to sustain and enhance the fabric of everyday life. Because we believe every surface is an opportunity. It’s what you’d expect from a pioneering and long-established paints company that’s dedicated to providing sustainable solutions and preserving the best of what we have today – while creating an even better tomorrow. Let’s paint the future together.

For more information please visit www.akzonobel.com

© 2024 Akzo Nobel N.V. All rights reserved.

Job Purpose

The Technology Integration Team develops and productizes reusable Color Technology Components that power AkzoNobel’s internal applications and external customer-facing tools, with focus areas in Colorimetric Engines/Building Blocks, Color Visualization, and AI/ML-enabled color services. The AI/ML Researcher owns the delivery of applied Machine Learning and statistical solutions for prioritized color and coatings use cases (e.g., formulation optimisation, pigment/raw material screening, quality prediction, and image-based inspection). The role translates business and R&D questions into well-structured datasets, validated ML models, and clear insights, and partners with engineering and product stakeholders to ensure solutions are deployment-ready, monitored, documented and usable in products and processes. Unlike generalist Researcher roles, this position brings dedicated applied ML/data science capability to the team, owning the end-to-end ML lifecycle from problem framing through deployment-ready model delivery.

Key Accountabilities

• Use case shaping & stakeholder partnering
• Engage with R&D, lab teams, manufacturing and product/application stakeholders to translate needs into well-defined ML problems, required inputs/outputs, constraints, and measurable success criteria.
• Lead or co-lead end-to-end delivery for assigned use cases from discovery and scoping through experimentation, validation, and implementation readiness.
• Communicate trade-offs (accuracy vs. explainability, latency, cost, robustness) and guide stakeholders toward fit-for-purpose solutions.
• Data preparation, analysis & modelling
• Acquire, clean, integrate, and analyze data from R&D/lab systems, manufacturing sources, and internal datasets; assess data quality, bias, leakage risk, and suitability for modelling.
• Develop, benchmark, and validate ML models (supervised/unsupervised) including baseline creation, feature engineering, model selection, hyperparameter tuning, and robustness checks.
• Domain application (color/coatings)
• Collaborate with color formulators, researchers, and lab teams to ensure modelling aligns with domain realities (measurement variability, protocols, material constraints).
• Define and drive dataset design/generation activities (e.g., DoE support, sampling strategy, labeling guidance, protocol awareness) to improve training data quality and coverage.
• Apply practical understanding of relevant color data types (spectral/Lab*/ΔE and variability) when designing targets, features, and evaluation strategies (as applicable).
• Computer vision (where relevant)
• Deliver image-based modelling tasks (e.g., defect detection, surface uniformity, pigment recognition) using classical CV and deep learning methods; define evaluation metrics and acceptance criteria aligned to operational deployment.
• Ensure peer review and alignment with Senior Researchers/architects for high impact use cases and model choices that may affect product/process behavior.
• Industrialization & production readiness
• Drive implementation readiness of assigned models: reproducible training/inference code, model artifacts, data dependencies, parameterization, testing approach, and clear runbooks.
• Contribute to MLOps practices as defined by the team: version control, experiment tracking, model registry inputs, monitoring requirements, alert thresholds, and retraining triggers.
• Partner with application/software teams to integrate model outputs into components/services (APIs, batch pipelines, or embedded usage) and support post-deployment stabilization.

Experience

Educational qualifications

• Master’s degree (preferred) or Bachelor’s degree in Data Science, Computer Science, Statistics, Applied Mathematics, Operational Research, Engineering or related quantitative field.

Relevant Experience

• Typically 5-7 years of experience applied machine learning / data science delivering solutions in industrial, manufacturing, R&D, engineering, or equivalent complex environments (or equivalent project portfolio demonstrating end-to-end ownership).
• Experience collaborating with cross-functional stakeholders (research, engineering, product) and delivering work in an agile/team setting.

Required Skills

• Strong Python skills for data analysis and ML (pandas, NumPy, scikit-learn; TensorFlow/ PyTorch beneficial).
• Solid understanding of ML fundamentals (regression, classification, clustering, model validation, overfitting control) and statistical reasoning.
• Working knowledge of SQL and working with structured datasets; o Familiarity with data pipelines or cloud platforms (Azure/Databricks) is a plus.
• Computer vision knowledge (OpenCV, CNNs/segmentation) is a plus depending on scope.
• Familiarity with Git and reproducible workflows; exposure to MLflow/ experiment tracking is desirable.
• Strong communication skills: ability to present findings clearly and document assumptions/limitations.
• Domain familiarity with color science (CIE color spaces, spectral data, ΔE) is advantageous but not mandatory.

At AkzoNobel we are highly committed to ensuring an inclusive and respectful workplace where all employees can be their best self. We strive to embrace diversity in a context of tolerance. Our talent acquisition process plays an integral part in this journey, as setting the foundations for a diverse environment. For this reason we train and educate on the implications of our Unconscious Bias in order for our TA and hiring managers to be mindful of them and take corrective actions when applicable. In our organization, all qualified applicants receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, age or disability.