职位描述

Description

  • This role sits within the client's Digital Practice and is aimed at a consultant-type data scientist — someone who can translate business problems into the language of data, and translate analytical output back into the language of the executive committee.
  • The successful candidate will diagnose client analytics maturity, structure ambiguous business issues into solvable analytical problems, design domain-informed features, build and tune models, and convert results into commercially meaningful recommendations. Delivery extends beyond analysis into service-isation — embedding models into dashboards, supporting automated AI-agent implementation, and using Gen AI and LLM techniques to deepen insight and compress analysis lead time. The role carries significant multi-stakeholder coordination across field operations, business functions, and IT.
  • Industry coverage spans consumer goods, commerce, retail, electronics, chemicals, and manufacturing.

工作要求

Requirement

  • Diagnose client analytical maturity — internal and external data utilisation, project pool, methodology and reporting quality — and establish To-Be direction through leading-practice benchmarking
  • Structure business issues into analytical problems solvable with data; document current state and improvement opportunities in structured reports
  • Design derived variables grounded in domain context, going beyond simple technical transformation; interpret data distribution and patterns to define and restructure segments
  • Assess limitations of internal data and integrate externally sourced and secondarily processed data into unified variables
  • Judge whether a single model is sufficient or whether hybrid approaches will generate synergy, and tune accordingly
  • Build and tune statistical, machine learning, and deep learning models, prioritising explainability over black-box approaches
  • Derive implications framed around business impact rather than accuracy alone, and communicate these in executive language
  • Apply Gen AI and LLM techniques to deepen insight and reduce analysis lead time
  • Service-ise analytical output through dashboard and model embedding, and support automated AI-agent implementation
  • Coordinate multi-layered stakeholders across field, business, and IT functions, including change management

Qualifications

  • Proficiency in data preprocessing and advanced statistical / machine learning modelling
  • Strong command of SQL and Python
  • Demonstrated use of descriptive, diagnostic, and predictive analytical methods
  • Hypothesis validation experience using statistical methods including regression analysis, ANOVA, and correlation analysis, alongside experience building ML/DL-based predictive models
  • Ability to design meaningful variables through combined understanding of business context and data
  • Capacity to read data structure and distribution against domain context and extract analytical value
  • Track record of improving model performance through variable redesign and weight / hyperparameter tuning, beyond straightforward application of existing models
  • Experience translating analytical results into business language for executive and operational audiences
  • Experience collaborating with and coordinating across multiple functions including business, IT, and DX
  • Experience structuring unstructured business problems into analytical tasks and identifying root causes
  • Experience combining qualitative insight with quantitative analysis to deliver tangible improvement

Preferred

  • Experience in consulting or an in-house strategy / analytics function
  • Bachelor's degree or above in Computer Science, Statistics, Industrial Engineering, or a related discipline
  • Industry data analytics experience in manufacturing, retail, commerce, or consumer goods
  • Practical application of Gen AI / LLM in analytical work — prompt engineering, RAG, analysis automation
  • Understanding of analytical data structures including DW/DM metric design
  • BI dashboard development experience (Tableau, Power BI, or equivalent)