Senior ML Engineer, GenAI for Smart Manufacturing

Micron Technology; Singapore; On-site; SGD 180,000 - 300,000; Full time

Job summary

Micron Technology in Singapore is seeking an ambitious Machine Learning Engineer (MTS) to drive large-scale model training and autonomous AI agents for manufacturing processes. You will collaborate with data scientists and engineers to deploy scalable AI/ML solutions across cloud environments and on-prem clusters.

The role requires deep knowledge of GPU architectures, distributed training, and LoRA/QLoRA fine-tuning, with a strong emphasis on PyTorch and CI/CD practices.

Qualifications

• Technical degree in CS or statistics; strong foundation in ML/AI.
• Experience ETL pipelines and big data processing at scale.
• Proven ability to deploy ML/AI solutions in cloud environments.

Responsibilities

• Architect and train large-scale models across multi-node GPU clusters.
• Optimize distributed training throughput and memory usage.
• Design autonomous AI agents for manufacturing workflows.
• Implement AI agent frameworks to orchestrate LLMs with internal APIs.
• Profile GPU performance and fix bottlenecks to maximize hardware use.
• Build data pipelines feeding ML models and GenAI apps.
• Develop data structures in Snowflake and Google Cloud for AI/ML.
• Create/maintain CI/CD pipelines for ML/AI solutions in the cloud.

Skills

GPU architecture

DDP

FSDP

LoRA/QLoRA

PyTorch

TensorRT-LLM

LangChain

LangGraph

Kubernetes

Python

CI/CD (Jenkins, Docker)

Big data processing

Education

Technical Degree in CS/Statistics or related field

Tools

PyTorch

Snowflake

Google Cloud

Nsight Systems

TensorRT

Job description

Micron Technology in Singapore is seeking an ambitious Machine Learning Engineer (MTS) to drive large-scale model training and autonomous AI agents for manufacturing processes. You will collaborate with data scientists and engineers to deploy scalable AI/ML solutions across cloud environments and on-prem clusters.

The role requires deep knowledge of GPU architectures, distributed training, and LoRA/QLoRA fine-tuning, with a strong emphasis on PyTorch and CI/CD practices.