Research Member of Technical Staff – Training Platform

About the role

Rhoda AI is building generalist intelligent robots and needs a Research Member of Technical Staff to own the training platform that powers model development. You will enable researchers to move from idea to result quickly by building orchestration, experiment management, and observability tools.

Key responsibilities

• Design and maintain training orchestration systems for large‑scale distributed model training on GPU clusters.
• Develop experiment management tooling for job configuration, tracking, reproducibility, and artifact handling.
• Create observability infrastructure to monitor loss curves, compute utilization, gradient statistics, and detect anomalies.
• Optimize the research iteration loop from experiment launch through results analysis.
• Manage job scheduling and cluster utilization to maximize GPU compute efficiency.
• Collaborate with training systems, data infrastructure, and research teams to support platform needs.

Required profile

• Strong software engineering background with experience in MLOps or ML platform engineering.
• Hands‑on experience with distributed training frameworks such as PyTorch DDP, FSDP, DeepSpeed, or Megatron.
• Proficiency in building experiment tracking, reproducibility, and artifact management solutions.
• Comfort operating GPU cluster environments using schedulers like Slurm or Kubernetes.
• Reliability engineering instincts, including monitoring, alerting, and failure recovery.

Required skills

• MLOps / ML platform engineering
• Distributed training frameworks (PyTorch DDP, FSDP, DeepSpeed, Megatron)
• Experiment tracking and artifact management
• GPU cluster management (Slurm, Kubernetes)
• Reliability engineering (monitoring, alerting, recovery)