Senior ML Engineer

We are looking for a Senior Data Engineer with a machine learning focus to join our data team, taking ownership of the data and feature pipelines that production ML systems depend on, training data preparation, feature engineering, and the path from a model that works in a notebook to one that runs, and keeps running, in production. You’ll work with a high degree of autonomy on the more complex and ambiguous problems across client engagements, partnering closely with data scientists and ML engineers, while mentoring junior engineers and helping to raise the bar on engineering practice. The foundations still matter: this is a data engineering role at heart, built on Python, PySpark, and platforms like Databricks, Snowflake, or AWS.

• Take ownership of designing, building, and maintaining the production data and feature pipelines that machine learning systems depend on, training data preparation, feature engineering, and reproducible, point-in-time-correct datasets, with minimal supervision
• Take models from prototype to production and keep them there: packaging and deployment alongside data science colleagues, model versioning, automated retraining, and the monitoring that catches drift and degradation before a client does
• Build and maintain the batch and ELT pipelines the ML work sits on, using Python, PySpark, and cloud data services (e.g., AWS Glue, EMR, S3, Redshift, Athena, or the Databricks / Snowflake equivalents), making sound decisions on architecture, data modelling, and tooling and weighing up trade-offs in performance, cost, and maintainability
• Own and continuously improve CI/CD for data and ML workflows, ensuring reliable, repeatable deployments of pipelines and models alike
• Mentor and support junior and mid-level data engineers, reviewing code and sharing knowledge to raise the standard of delivery across the team
• Troubleshoot and resolve complex issues in live data pipelines and deployed models, acting as a point of escalation for the wider team
• Partner with data scientists, cross-functional teams, and client stakeholders to understand data and ML requirements and shape them into technical solutions

• Have hands-on experience building the data and feature pipelines behind production machine learning, training data preparation, feature engineering, and feature stores, and understand why reproducibility and point-in-time correctness matter more than they first appear
• Have worked with MLOps in practice rather than in principle: model versioning and registries, experiment tracking, automated retraining, and monitoring for drift and model degradation (MLflow, SageMaker, Databricks ML, Vertex AI or comparable)
• Are comfortable with ML frameworks such as scikit-learn, PySpark MLlib, or PyTorch / TensorFlow, enough to productionise a data scientist’s work, reason about what a model needs from the data, and spot where a training pipeline and a serving pipeline have quietly diverged. You don’t need to be the person designing the model
• Have designed and operated production data pipelines on a modern cloud data platform, AWS (Glue, EMR, Redshift, S3/Lake Formation), Databricks, Snowflake or comparable, and can talk about which of those decisions would travel to a different one, and which wouldn’t
• Have deep SQL skills and hands-on experience with data modelling (e.g., Kimball, medallion/lakehouse patterns, or domain-driven modelling), this is the load-bearing, provider-agnostic skill underneath almost all of our pipeline work
• Have strong Python and PySpark for batch processing, with production-grade, well-tested code. The bulk of this work is batch and ELT, but you’ll need to have experienced event-driven or micro-batch pipelines and know where the design assumptions stop holding
• Have owned orchestration in production, dependency design, retries, SLAs, and the observability that tells you a pipeline failed before the client does
• Have experience owning the delivery of data engineering work with limited oversight, including CI/CD management and support
• Have experience mentoring or supporting the development of less experienced engineers
• Are comfortable working directly with stakeholders to shape technical solutions from ambiguous requirements, presenting your work, estimation and discussing tradeoffs and risks.