Robotics Deployment Engineer: VLA/WAM & Real‑World Ops

Flexion Robotics; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Benefits offered by this job

Competitive compensation

Relocation sponsorship

Open PTO policy

Health, dental & vision coverage

Open/transparent culture

Job summary

Flexion Robotics in San Francisco is seeking a VLA/WAM Integration and Deployment specialist to turn VLA and World-Action Model research into reliable capabilities on humanoid robots.

You will deploy across the full loop, optimize inference and latency, and connect policies with higher‑level agents, gathering real‑world data to drive improvements.

This hands‑on role suits someone who can blend engineering, data pipelines, and robotics hardware to deliver robust autonomous behavior.

Qualifications

• Strong Python and PyTorch skills with modern DL systems.
• Experience deploying learning-based models in robotics or real-time systems.
• Familiarity with VLA, VLM, diffusion, or multimodal architectures.
• Systems intuition around latency, serving, data pipelines, logging and debugging.
• Experience designing robust evaluation and diagnosing failures.
• Comfort working directly with robots and real-world hardware.
• Ability to bridge training, deployment, data, and infrastructure.

Responsibilities

• VLA/WAM Deployment: deploy and validate policies across humanoid platforms.
• Fast Inference: optimize serving latency, action generation and closed-loop control.
• Agentic Integration: connect policies with higher-level agents and replanning.
• In‑Situ Data Collection: improvements for teleoperation, rollouts, data collection.
• Evaluation & Testing: create reproducible real-world evaluations and comparisons.
• Deployment‑to‑Training Loop: turn deployments into data for pre/post-training.

Skills

Python

PyTorch

VLA/VLM

Model deployment

Robotics systems

Latency optimization

Evaluation design

Tools

TensorRT

torch.compile

Model compression

Job description

Flexion Robotics in San Francisco is seeking a VLA/WAM Integration and Deployment specialist to turn VLA and World-Action Model research into reliable capabilities on humanoid robots.

You will deploy across the full loop, optimize inference and latency, and connect policies with higher‑level agents, gathering real‑world data to drive improvements.

This hands‑on role suits someone who can blend engineering, data pipelines, and robotics hardware to deliver robust autonomous behavior.