Autonomy Engineer — VLA Pre-Training

Humanoid; San Diego (CA); On-site; USD 180,000 - 300,000; Full time

Benefits offered by this job

Health insurance

PTO 23 days + holidays

401(k) matching

Equity

Free daily lunch

Collaborative team

Ownership of initiatives

Job summary

Humanoid seeks an Autonomy Engineer focused on VLA pre‑training to advance the world’s most capable humanoid platform. You will pre‑train base models on a diverse corpus, fine‑tune policies for task performance, and shape data collection and synthetic data generation workflows.

The role emphasizes deep learning, with or without robotics background, and collaboration with MLOps and data teams to scale training and optimize edge inference. California base salary is competitive.

Qualifications

• 3+ years building deep‑learning systems with shipped models or publications.
• Hands‑on experience with LLMs, VLMs, or image/video generative models.
• Experience with DL infrastructure: streaming data, checkpointing, distributed training.
• Strong Python + PyTorch/JAX; able to profile, debug numerics, and write research code.
• Familiar with modern software engineering practices.
• Able to document experiments and communicate trade‑offs clearly.

Responsibilities

• Post‑train policies via behavior cloning and RL; own data‑to‑deployment loop.
• Partner with Data Collection to define good data, identify failures, ensure diversity.
• Collaborate with external partners for pretraining data supply.
• Run pre-/mid-/post‑training on VLA stack; explore new modalities and architectures.
• Build and maintain pipelines: ingest synthetic data, version data, label, curate datasets.
• Work with MLOps & Data Platform to scale distributed training and edge inference.

Skills

Python

PyTorch

JAX

Deep learning

Distributed training

Data pipelines

MLOps

Tools

OpenVLA

Job description

Humanoid seeks an Autonomy Engineer focused on VLA pre‑training to advance the world’s most capable humanoid platform. You will pre‑train base models on a diverse corpus, fine‑tune policies for task performance, and shape data collection and synthetic data generation workflows.

The role emphasizes deep learning, with or without robotics background, and collaboration with MLOps and data teams to scale training and optimize edge inference. California base salary is competitive.