STMicroelectronics

Manufacturing Data Science - IPP - Expert System

STMicroelectronics
Integrated Device ManufacturingSingapore, SingaporeOnsitePosted 1 month ago

About the role

AI summarised

Design and build agentic AI-driven expert systems to enable predictive and autonomous decision-making within complex manufacturing environments. This role involves developing advanced AI agents, leveraging simulation/digital twins, and integrating predictive models to optimize production flows.

IDMOnsiteData Science

Key Responsibilities

  • Develop agent-based and multiagent AI systems capable of planning, deciding, and acting on behalf of production assets.
  • Design reasoning and decision policies utilizing rules, optimization, and machine learning to manage real-time manufacturing scenarios (e.g., dispatching, scheduling, quality control).
  • Utilize simulation and digital twins to model production flows, equipment behavior, and system constraints for safe strategy testing.
  • Integrate predictive models (e.g., machine health, yield) into expert systems to support proactive and prescriptive operational decisions.
  • Evaluate AI solutions against KPIs including throughput, WIP, cycle time, OEE, and cost, driving iterative performance improvements.
  • Document system architecture, algorithms, experiments, and support IP creation related to agentic AI systems.

Requirements

  • Bachelor's or Master’s degree in Computer Science, AI, Data Science, Operations Research, or relevant Engineering field.
  • Strong foundation in AI and Machine Learning principles.
  • Exposure to Reinforcement Learning or Multiagent RL is required.
  • Experience with planning and decision-making under uncertainty.
  • Proficiency in Python and AI/ML ecosystems (e.g., PyTorch, TensorFlow, scikit-learn).
  • Familiarity with simulation or modeling of manufacturing/logistics systems (e.g., digital twins).
  • Strong analytical skills to formalize complex shopfloor logic into constraints and objectives.
  • Ability to work with real operational data, including cleaning and feature engineering.