Staff Multimodal ML Engineer for Autonomous Robots

Watney; San Francisco (CA); On-site; USD 140,000 - 200,000; Full time

Job summary

Watney is building autonomous robotic systems to accelerate data center buildouts. The ML Engineer will turn live robot fleets into robust models, curate large-scale data, and evaluate deployments across perception, planning, navigation, and manipulation.

You will train and refine models with multimodal sensor data and help ensure low-latency inference for a live fleet. The role requires hands-on development in Python and ML frameworks, with deployment across production environments and

Qualifications

• Experience building multimodal models on GPUs with large video data.
• End-to-end dataset handling: collection, labeling, and curation.
• Production training and inference pipelines in Python (PyTorch/JAX/Triton).
• Strong debugging skills with real-world model performance evidence.
• Familiarity with cluster management systems (Ray/Slurm/Kubernetes).

Responsibilities

• Own model quality end-to-end: dataset curation, training, evaluation.
• Work with cameras, IMUs, LiDAR, and odometry data to shape training sets.
• Train and evaluate policies; run data/feature/architecture ablations.
• Deploy results to production with low-latency, 24/7 inference.
• Collaborate with Teleoperations and Fleet Operations to obtain data.

Skills

Python

PyTorch

JAX

Triton

Multimodal data

Robot perception

Tools

Ray

Slurm

Kubernetes

Git

Job description

Watney is building autonomous robotic systems to accelerate data center buildouts. The ML Engineer will turn live robot fleets into robust models, curate large-scale data, and evaluate deployments across perception, planning, navigation, and manipulation.

You will train and refine models with multimodal sensor data and help ensure low-latency inference for a live fleet. The role requires hands-on development in Python and ML frameworks, with deployment across production environments and