VLA Pre-training Engineer (Deep Learning)

TL;DR

VLA Pre-training Engineer (Deep Learning): Developing and training vision-language-action policies for hirify.global robots with an accent on pre-training base models on multi-embodiment datasets and fine-tuning for specific tasks. Focus on building continuous data pipelines, scaling distributed training, and optimizing models for real-time edge inference.

Location: On-site in London, UK

Company

A robotics company building commercially-scalable and safe hirify.global robots to amplify human potential.

What you will do

• Post-train policies using behavior cloning and reinforcement learning, managing the full loop from data to deployment.
• Collaborate with Data Collection teams to define data requirements, identify failure modes, and ensure dataset diversity.
• Execute pre-, mid-, and post-training on the VLA stack, exploring new modalities and architecture changes.
• Develop and maintain pipelines for ingesting synthetic data and teleop logs with weak-supervision labeling.
• Partner with MLOps and Data Platform teams to scale distributed training and optimize for edge inference.

Requirements

• Must be based in or able to work on-site in London, UK
• 3+ years of experience building deep-learning systems with shipped models or published research.
• Hands-on experience with LLMs, VLMs, or image/video generative models (architecture, training, and inference).
• Proficiency in Python and PyTorch/JAX, including profiling and debugging numerics.
• Experience with deep learning infrastructure, including streaming datasets and distributed training strategies.
• Strong software engineering practices and ability to document experiments clearly.

Nice to have

• Experience in robotics or autonomous driving.
• Application of Reinforcement Learning (RL) to LLMs or robotics.
• Familiarity with VLA models and frameworks like OpenVLA or Physical Intelligence (π).
• Proven record of productizing deep networks with latency and throughput constraints.
• Publications at top-tier DL conferences or significant open-source contributions.

Culture & Benefits

• Competitive equity and stock options.
• Over 30 paid days off, including annual leave and company closure days.
• Private healthcare including virtual and in-person care.
• Pension scheme with employer contributions.
• Daily catered breakfast, lunch, and snacks in the office.
• Direct access to founding leadership and ownership of key initiatives.