Micron Technology - Senior Data Scientist
Building agentic AI products, ML services, and decision-support platforms for advanced memory manufacturing.
-
AI Platform / Agentic Systems Lead - Internal AI Product
Led the design and delivery of a production-grade agentic AI system used by engineers across multiple global sites to diagnose complex operational issues.
Built an LLM-driven assistant combining retrieval, tool orchestration, and reasoning workflows to support decision-making in high-ambiguity scenarios.
Owned the system end-to-end: problem definition, architecture design, gate-checked evaluation loops, deployment, and adoption.
Designed production checks for retrieval quality, tool-call behaviour, groundedness, and domain review before release.
Resulted in ~509 engineering hours saved per week and ~50% reduction in time-to-diagnosis, with sustained usage across sites.
-
Agentic AI Platform Leadership
Led the development of Micron's first production agentic AI system, establishing architecture patterns, use-case scoping, and best practices for enterprise-wide agentic AI adoption.
Designed and implemented an LLM-driven agent framework combining retrieval, reasoning, and tool orchestration to support complex diagnostic workflows.
Acted as technical lead and primary author, shaping standards later reused across multiple AI initiatives.
Recognised with the 2025 Micron Data Science Technical Seminar Innovation Award.
-
CVD Tool Qualification Reduction - Data Science with Business Impact
Led an applied ML and analytics initiative to optimise equipment qualification workflows, reducing end-to-end cycle time by ~70%.
Built feature engineering and modelling pipelines to identify critical qualification signals, enabling earlier decision-making and faster tool readiness.
Reduced qualification cycles from an average of 7 runs to 2, delivering ~US$400K annual savings through increased equipment availability.
-
Global Diagnostic ML Pipeline - Internal AI Platform Adoption
Designed and scaled reusable diagnostic modelling pipelines adopted by global data science teams for rapid issue diagnostics.
Abstracted common analysis patterns into a self-service framework, reducing time-to-insight for exploratory investigations and improving consistency across teams.
Served as technical owner and primary author, driving internal adoption and knowledge sharing across sites worldwide.
-
FOUP Gasket Warping Detection - Computer Vision in Production
Led the delivery of a production computer vision system to detect mechanical defects impacting downstream quality.
Designed, trained, and deployed a vision-based detection pipeline integrated into operational workflows, enabling early intervention and yield protection.
Achieved a baseline ~0.4% yield loss avoidance, with the system running in live production environments.
