AI Infrastructure Engineer

Matter Inc • Menlo Park, CA

21d

About The Position

Matter is building an AI-native autonomy stack for physical manufacturing in the United States. Its MatterOS platform is a unified software layer for factory operations, process orchestration, and autonomy deployment. The AI Infrastructure Engineer will build and own the data pipeline connecting factory-floor machine telemetry to cloud-based model training, supporting MatterOS and its Sim2Real training pipeline.

The role involves designing and operating systems across edge computing, cloud training infrastructure, and data pipelines. The engineer will ensure data is tagged with machine identity, process state, and production context so it can be used to train AI models, and will build infrastructure from scratch for autonomous physical manufacturing.

Requirements

• At least 3 years of experience in ML infrastructure, MLOps, or data engineering in a production environment.
• Strong command of distributed data systems such as Kafka or Flink, or equivalent, and time-series databases such as InfluxDB or TimescaleDB.
• Experience managing GPU clusters and distributed training using SLURM, Ray, or Kubernetes-based systems.
• Familiarity with industrial protocols such as OPC UA, MQTT, or Modbus, or willingness to learn quickly.
• Proficiency in Python.
• Systems thinking and understanding that data quality, rather than data volume, is critical to making AI work in constrained physical environments.

Nice To Haves

• Comfort with C++ or Rust for performance-critical edge components.
• Experience with NVIDIA Isaac Sim, ROS2, or edge AI deployment on Jetson, FPGA, or similar hardware.
• Background in industrial IoT or factory automation systems.
• Familiarity with model serving frameworks such as Triton, TorchServe, or ONNX Runtime.

Responsibilities

• Design and maintain an edge-to-cloud data pipeline that preserves semantic context end-to-end.
• Build and manage GPU compute infrastructure for VLA model training, experiment tracking, and distributed training workflows.
• Implement a data collection layer for complete capture from modular assembly workcells, including camera feeds, sensor streams, machine state, and process metadata.
• Develop feature-engineering pipelines that transform raw operational data into structured training inputs for AI models.
• Manage model deployment to factory edge hardware, including latency, versioning, rollback, and production monitoring.
• Build observability systems to surface model performance degradation, data drift, and equipment anomalies in real time.
• Collaborate with AI researchers to translate model requirements into infrastructure specifications and vice versa.

What This Job Offers

• Job Type: Full-time
• Education Level: No Education Listed