Senior ML Infrastructure Engineer — GPU & Robotics

Nimble; San Francisco (CA); On-site; USD 210,000 - 300,000; Full time

Benefits offered by this job

Unlimited Flexible Time Off

Health Insurance

Paid Parental Leave

Commuter Benefits

Referral Bonus

401k

Equity

Job summary

Nimble is seeking a Software Engineer to join the ML Infrastructure team. You will help build the training and inference systems that power our general-purpose warehouse robots.

You’ll own training infrastructure end to end, keeping GPUs highly utilized and enabling researchers to launch new experiments with a single command. You’ll collaborate with ML and Robotics teams to design scalable, high-throughput platforms for model development.

Qualifications

• BA/BS/MS/PhD in CS or related field, or equivalent practical experience.
• 4+ years in infrastructure, distributed systems, ML systems, robotics, or related area.
• Experience with Rust, Go, Python, or C++.
• Experience with PyTorch or JAX.
• Strong understanding of distributed systems and performance optimization.
• Experience with Kubernetes orchestration and large distributed jobs.
• Ability to optimize GPU memory hierarchy and multi-GPU operations.

Responsibilities

• Design, develop, and maintain ML training infrastructure enabling AI teams to run training jobs efficiently.
• Build low-latency inference pipelines for production robotics workloads.
• Develop, tune, and optimize CUDA kernels.
• Design scalable training-platform systems for high-throughput data ingestion and distributed training.
• Participate in design reviews and evaluate technical tradeoffs.

Skills

Rust

Go

Python

C++

PyTorch

JAX

Distributed systems

Kubernetes

GPU memory tuning

Systems optimization

Cross-functional collaboration

Education

Bachelor's, Master's, or PhD in Computer Science or related field

Tools

Kubernetes

PyTorch

JAX

DeepSpeed

Megatron

NCCL

Parquet

Arrow

Job description

Nimble is seeking a Software Engineer to join the ML Infrastructure team. You will help build the training and inference systems that power our general-purpose warehouse robots.

You’ll own training infrastructure end to end, keeping GPUs highly utilized and enabling researchers to launch new experiments with a single command. You’ll collaborate with ML and Robotics teams to design scalable, high-throughput platforms for model development.