Humanoid

VLA Pre-training Engineer - Deep Learning

Humanoid
Apply
2 months ago
London, United KingdomMid Level

Responsibilities

  • Post-train policies using behavior cloning and reinforcement learning, owning the process from data through deployment.
  • Partner with data collection teams to define quality standards, identify failure modes, and ensure data diversity and coverage.
  • Coordinate with external partners to maintain a supply of high-quality pretraining-scale data.
  • Run pre-training, mid-training, and post-training on the VLA stack while exploring new modalities and architecture changes.
  • Build and maintain pipelines for synthetic data and teleoperation logs, including versioning, weak-supervision labeling, dataset curation, and automated failure-case discovery.
  • Collaborate with MLOps and Data Platform teams to scale distributed training and optimize models for real-time edge inference.

Requirements

  • 3+ years of experience building deep-learning systems in industry or research, with shipped models or published artifacts.
  • Deep hands-on experience with at least one of LLMs, VLMs, or image/video generative models, including architecture, training, and inference.
  • Experience with streaming datasets, checkpointing and state management, and distributed training strategies.
  • Strong Python and PyTorch or JAX skills, including profiling, numerical debugging, and maintainable research code.
  • Familiarity with modern software engineering practices and the ability to document experiments and communicate trade-offs clearly.
  • Robotics or autonomous driving experience, reinforcement learning for LLMs or robotics, VLA model experience, deep-net productization, relevant publications, open-source contributions, or familiarity with OpenVLA and Physical Intelligence models are preferred.

Benefits

  • Competitive equity through stock options with meaningful upside.
  • 30+ paid days off, including 23 days of annual leave, UK bank holidays, and additional company closure days.
  • Private healthcare with virtual and in-person care.
  • Pension scheme with an 8% total contribution, comprising 5% employee and 3% employer contributions, on full earnings.
  • Free daily breakfast, catered lunch, and snacks in the office.
  • London-based role with in-office collaboration and access to founding leadership and product ownership.

Tech Stack

Categories

Humanoid

About Humanoid

201-500 employees
Contact me