Data Intelligence Machine Learning Engineer

About Dyson

Dyson is committed to relentless innovation, pushing the limits within engineering, artificial intelligence, and robotics. Our Data Intelligence team is central to this mission, responsible for shaping the company's future through smart data strategies and pipelines that power the next generation of connected products.

You will collaborate with Dyson's global engineering experts and external technology partners in a dynamic environment focused on innovation, execution, and meaningful impact.

Role Overview

We seek a Data Intelligence Machine Learning Engineer to build and implement proprietary tools that automate the data annotation process. The goal is to minimize manual labeling efforts by using advanced techniques such as Active Learning, Weak Supervision, and Synthetic Data Generation. This role acts as the critical link converting raw data into accurately labeled datasets at scale.

Key Responsibilities

• Design and deploy end-to-end automated label pipelines using frameworks like Snorkel, Cleanlab, or custom active learning methods.
• Create "Human-in-the-Loop" systems enabling model pre-labeling with human review reserved for uncertain data points.
• Develop algorithmic quality assurance mechanisms to detect and correct mislabeled or noisy data.
• Partner with software engineers to integrate labeling tools with data lakes and machine learning training systems.
• Optimize teacher models to produce high-quality pseudo-labels that improve student model training.
• Establish and maintain scalable data preparation workflows optimized for data quality, efficiency, and integration into MLOps pipelines.
• Perform sophisticated data visualization and analysis using feature engineering to translate raw data into actionable insights supporting research and product deployment.
• Collaborate closely with Data Scientists, Software Engineers, and Product teams to uphold data accuracy and usability.

Candidate Profile

• Minimum three years of experience in machine learning engineering focused on data-centric AI applications such as computer vision or natural language processing pipelines.
• Expert-level proficiency in Python and the machine learning ecosystem (PyTorch or TensorFlow, NumPy, Pandas, Scikit-learn).
• Proven expertise in automated labeling approaches including Weak Supervision (using labeling functions) and Active Learning techniques (uncertainty and diversity sampling).
• Experience with both SQL and NoSQL databases, managing large unstructured data types like images, texts, or audio.
• Familiar with cloud-based labelling solutions like AWS SageMaker Ground Truth, Google Cloud Vertex AI, or Azure ML.
• Experience in data version control systems such as DVC to manage dataset versions.
• Demonstrated history of developing auto-labeling systems or handling large-scale data annotation workflows.
• Strong background in building scalable, maintainable data pipelines including cleansing, transformation, and storage in cloud or hybrid environments.
• Advanced skills in feature engineering, data analysis, and visualization tools including Jupyter Notebooks, Tableau, and Power BI.
• Excellent communication skills allowing clear documentation and smooth cooperation with technical and non-technical stakeholders.
• Ability to work efficiently balancing speed and quality in a rapidly changing environment while staying current on technological advances.
• Bachelor's or Master's degree in Computer Science, Engineering, Mathematics, Data Science, or a related field.

Diversity and Inclusion Statement

Dyson is an equal opportunity workplace valuing diverse perspectives, encouraging applications from all backgrounds without discrimination based on race, color, religion, sex, sexual orientation, gender identity, age, disability, or veteran status.