Forecast ML Scientist: Pipelines & Dynamic Reforecast

HM Revenue & Customs; United Kingdom; On-site; GBP 70,000 - 120,000; Full time

Job summary

HM Revenue & Customs seeks an experienced ML Engineer to improve yield and forecast accuracy using data-driven models. You will develop multivariate predictive ML models learning from past data across financial and non-financial variables to forecast KPIs and support assumption management for non-reported items.

The role includes dynamic re-forecasting in the Decision Execution module and building scalable deployment pipelines in production using AWS, Sagemaker, PyTorch and TensorFlow.

Qualifications

• Solid foundations in ML modelling, comfortable with a range of supervised and unsupervised ML algorithms.
• Ability to design and maintain ML pipelines and architecture on modern frameworks to meet functional business requirements.
• Hands on experience in using cloud technologies (AWS) in a production environment, as well as ML platforms (e.g. AWS Sagemaker).
• Knowledge and experience of deep learning frameworks (e.g. PyTorch, Tensorflow).
• Proven ability to build ML model deployment pipelines in a scalable production environment.

Responsibilities

• Develop multivariate predictive ML models to forecast KPIs such as Yield and Strategic Objectives.
• Implement dynamic re-forecasting in response to decisions in the Decision Execution module.
• Build robust ML pipelines and cloud-based production workflows to support forecasting.

Skills

ML modelling

ML pipelines

Deployment pipelines

Cloud platforms

Tools

AWS

SageMaker

PyTorch

TensorFlow

Job description

HM Revenue & Customs seeks an experienced ML Engineer to improve yield and forecast accuracy using data-driven models. You will develop multivariate predictive ML models learning from past data across financial and non-financial variables to forecast KPIs and support assumption management for non-reported items.

The role includes dynamic re-forecasting in the Decision Execution module and building scalable deployment pipelines in production using AWS, Sagemaker, PyTorch and TensorFlow.