Machine Learning Engineer

Role Purpose

The Machine Learning Engineer will lead the AI and ML roadmap for the Digital Innovation & Product Hub, designing and shipping production-grade machine learning systems that power the Media teams at dentsu. From building robust training and inference pipelines, to defining how we evaluate model quality and business impact, this role owns the full ML lifecycle, partnering with Product Managers, specialism leads, data engineers and developers to turn business problems into deployed, monitored, well-evaluated models that drive measurable outcomes for our clients.

Core Accountabilities

• Leading the AI and ML roadmap for the team, identifying high-value opportunities, prioritising against business impact, and translating strategic goals into a clear, sequenced plan of ML initiatives
• Designing, building and maintaining end-to-end ML pipelines covering data ingestion, feature engineering, training, validation, deployment and retraining - with reproducibility, scalability and observability baked in
• Owning model evaluation - defining offline and online metrics, building eval sets, running A/B tests and validating models for accuracy, fairness, robustness and business impact before and after deployment
• Establishing MLOps best practices across the team - experiment tracking, model registry, versioning, CI/CD for models, and infrastructure-as-code - alongside clear technical documentation
• Monitoring models in production, detecting drift, debugging performance regressions, and iterating to keep latency, cost and accuracy within agreed thresholds
• Partnering with developers, data engineers and Product Managers to expose models via well-designed APIs, and working with specialism leads to embed ML capabilities into Media team workflows
• Collaborating with our internal Security and Legal teams to ensure models comply with dentsu’s Security Policies, data handling standards and responsible-AI principles

Skills

We’re keen to meet anyone who’s comfortable with most of the below and are flexible if you have strengths or weaknesses in particular areas.

Technical

• Strong experience training, fine-tuning and deploying machine learning models in production, with a solid grounding in classical ML and modern deep learning
• Strong Python skills, with hands-on experience using ML frameworks such as PyTorch, TensorFlow, scikit-learn and Hugging Face, plus working with SQL and large-scale data tooling
• Experience building ML pipelines and MLOps tooling - e.g. Airflow, Kubeflow, MLflow, Weights & Biases, Vertex AI or SageMaker - and deploying models on cloud (GCP ideally)
• Well versed in agile methodologies, Git and version control best practices
• Deep experience with model evaluation - offline metrics, eval set design, A/B testing, drift detection, fairness checks and validating models against business KPIs
• Comfortable with Docker, containerised model serving and exposing models via APIs for downstream developers and applications
• Exposure to LLMs, RAG or generative AI is a bonus, but not essential