Autonomy Engineer: VLA Pre-Training & Edge AI Pipelines

Humanoid; United States; On-site; USD 180,000 - 300,000; Full time

Benefits offered by this job

Health insurance

PTO 23 days

401(k)

Equity

Free lunch

Collaborative environment

Product ownership

Job summary

Humanoid in the United States is hiring an Autonomy Engineer focused on VLA pre-training. You will pre-train base models on diverse trajectories, fine-tune policies for specific tasks, and shape data collection for robust synthetic data generation with a strong emphasis on deep learning.

You will collaborate with Data Collection and external partners to ensure high-quality pretraining data, run training cycles on the VLA stack, and optimize pipelines for edge deployment while leveraging Python,

Qualifications

• 3+ years building deep-learning systems with shipped models or published artifacts.
• Experience with at least one: LLMs, VLMs, or image/video generative models (architecture, training, inference).
• Experience with deep learning infrastructure: streaming datasets, checkpointing, state management, distributed training strategies.

Responsibilities

• Post-train policies via behavior cloning and RL; own the data-to-deployment loop.
• Collaborate with Data Collection to define good data and coverage.
• Coordinate with external partners to maintain pretraining data supply.
• Run pre-/mid-/post-training on VLA stack and experiment with modalities.
• Maintain pipelines: ingest synthetic data, version datasets, label data, surface failures for retraining.
• Work with MLOps to scale distributed training for real-time edge inference.

Skills

Deep learning systems

LLMs / VLMs / generative models

Python programming

PyTorch / JAX

Tools

PyTorch

JAX

Distributed training

Job description

Humanoid in the United States is hiring an Autonomy Engineer focused on VLA pre-training. You will pre-train base models on diverse trajectories, fine-tune policies for specific tasks, and shape data collection for robust synthetic data generation with a strong emphasis on deep learning.

You will collaborate with Data Collection and external partners to ensure high-quality pretraining data, run training cycles on the VLA stack, and optimize pipelines for edge deployment while leveraging Python,