Research Engineer

The role

You will own training and inference infrastructure end to end. You will scale distributed training, build research codebases, design data pipelines, and ship low-latency inference for real-time robot control.

What you will do

• Scale distributed training across a multi-cloud GPU fleet using sharding, activation checkpointing, and memory optimization
• Build a research codebase and scheduling system built for fast iteration, automatic retries, and painless failure recovery
• Design high-throughput pipelines for terabytes of multimodal robot data
• Ship low-latency inference for real-time robot control using quantization, distillation, and compilation
• Profile GPU utilization, I/O bottlenecks, and memory fragmentation

Required qualifications

• 7+ years of engineering experience, including leading projects in HPC or ML infrastructure
• PyTorch and distributed training (DeepSpeed, Accelerate, FSDP), including mixed precision and gradient accumulation
• Hands-on experience with cloud GPU environments (GCP/AWS) and Kubernetes
• Understanding of distributed systems at a low level: race conditions, memory management, NCCL and inter-node communication
• Ability to design, build, and operate systems end to end

Nice to have

• Robotics data formats (MCAP, Protobuf) or multimodal models such as VLAs
• Custom kernels (Triton), compilers, or runtime optimization
• Experience as a founding or early infrastructure hire
• Side projects involving robots, simulators, or hardware

About the company

Dyna Robotics builds general-purpose robots powered by a proprietary embodied AI foundation model. The team comes from Google DeepMind, Meta, and Cruise.