Data Science Engineer

Job Description

Role: ML/ Data Science Engineer

Rate: £500 - £550 Inside IR35

Location: Central London, hybrid 3 days per week on-site

Duration: 6 months initial contract

About the role

We're looking for a hands-on ML / Data Science Engineer to design, build and optimise production ML models on AWS.

You'll work across the full model lifecycle - from choosing the right algorithm through to tuning, deployment and monitoring - using classical ML techniques.

This role will suit someone who can explain not just what they built, but why: which models they chose, the trade-offs, and how they measured success in production.

Key responsibilities

• Design, build and validate classical ML models (e.g. regression, tree-based ensembles, clustering) to solve real business problems
• Select appropriate algorithms and clearly justify model choices to technical and non-technical stakeholders
• Develop clean, well-tested, production-ready Python code
• Build, train and deploy models on AWS using Amazon SageMaker and related services
• Tune model performance through hyperparameter optimisation, feature engineering and error analysis
• Instrument models in production, defining and tracking MLOps metrics such as accuracy, drift, latency and throughput
• Collaborate with data engineers, data scientists and product teams to take models from prototype to production

Essential skills & experience

• Classical / traditional ML: Proven experience with models such as linear and logistic regression, decision trees, random forests, gradient boosting (XGBoost, LightGBM), SVMs and clustering. You can explain which models you've used, why you chose them and what alternatives you considered.
• Python: Strong proficiency in Python and its data/ML ecosystem (pandas, NumPy, scikit-learn), writing maintainable, production-quality code.
• AWS: Hands-on experience building, training and deploying ML models on the AWS stack, ideally with experience in Amazon SageMaker
• Model tuning & MLOps: Practical experience in hyperparameter tuning and performance optimisation. You can articulate the metrics you'd track for MLOps (e.g. model accuracy, precision/recall, data and concept drift, latency, cost) and how you've instrumented models to monitor and improve performance in production.