A twelve month contract to take several models from promising notebook results to something running reliably in production, with monitoring and a retraining path. We are being deliberate about this hire. The problems are document extraction, demand forecasting and anomaly detection on operational data, all with real business consequences attached.
Contract engagement with a solid possibility of conversion to a permanent role for the right person.
Responsibilities
- Productionising models as versioned, monitored services
- Feature pipelines and the training data that feeds them
- Evaluation methodology, including what the model is allowed to be wrong about
- Working with domain experts to define what good actually means
Requirements
- Four or more years applying machine learning to real problems, not only research
- Strong Python, with PyTorch or TensorFlow and the usual scientific stack
- MLOps experience covering deployment, drift monitoring and retraining
- NLP or document understanding experience is directly relevant
- The honesty to report that a model is not good enough yet
Disclaimer: This job posting has been aggregated from external source. Role details, content, and availability are subject to change. Applicants are advised to confirm the latest information directly on the company website before applying.