Responsibilities

• Design and implement end-to-end ML pipelines from data ingestion through model deployment and signal delivery
• Transform client-specific Jupyter notebooks into modular, config-driven pipelines using orchestration tools such as Prefect/Airflow
• Build robust API connectors handling schema evolution, incremental updates, and data quality validation
• Implement comprehensive machine learning model evaluation frameworks blending technical metrics (precision, recall, PRAUC, probability calibration) with business outcomes
• Develop AutoML capabilities optimised for time-series behavioural data and subscription lifecycles
• Implement sophisticated feature engineering for event-based data
• Design multi-model systems handling various prediction horizons and conversion definitions
• Optimise hyperparameter tuning using frameworks like Optuna, AutoGluon, or H2O
• Establish MLOps practices appropriate for a small team: experiment tracking, model registry, and monitoring
• Collaborate with engineering on CI/CD pipelines, testing frameworks, and deployment automation

Requirements

• 5-8+ years building production ML systems with demonstrable business impact
• Strong experience with time-series analysis and behavioural event modelling
• Deep expertise in Python with high code quality standards
• Experience with modern ML stack (e.g. pandas/polars, sklearn, xgboost, PyTorch/TensorFlow)
• Proven track record delivering end-to-end ML pipelines: ingestion → feature engineering → training → deployment → monitoring
• Hands-on experience with cloud data warehouses (e.g. BigQuery, Snowflake)
• Track record of building automated, scalable systems from initial prototypes
• Right to Work in the UK (we cannot sponsor visas)
• Ability to work from Central London office 3 days/week (we believe in-person collaboration is crucial at this early stage)
• You may be a great fit if you have any of the following...

Preferred

• AutoML framework experience (e.g. AutoGluon, TPOT, Optuna, H2O.ai)
• MLOps tooling (e.g. MLflow, Weights & Biases, Evidently)
• Hands-on experience with orchestration tools (e.g. Prefect, Airflow, Dagster)
• Building robust API/ETL connectors with retry logic and incremental loading
• Statistical depth beyond standard metrics: calibration, cost-sensitive learning, causal inference
• Passionate about leveraging the latest LLM tooling for accelerated AI-enhanced delivery without compromising on quality
• Domain knowledge bonus points - Marketing attribution and conversion modelling, Mobile app analytics and user lifecycle prediction, Ad-tech ecosystem and privacy regulations (ATT, GDPR), Subscription business metrics and retention modelling

Benefits

• Holidays: 25 days of annual paid leave, plus bank holidays.
• Flexibility: 3 days/week in Central London office, remote otherwise
• Equipment: Top-spec MacBook Pro and any tools you need
• Learning Budget: Conferences, courses, and resources to stay current