Requirements

Hands-on experience post-training/fine-tuning large-scale models (language, video, or other foundation models)
Experience with world models, autoregressive generation, and long-horizon generation
Experience with diffusion/flow models and understanding of 3D vision
Strong understanding of model architecture and the ability to contribute meaningfully to architectural/training decisions
Strong hands-on engineering skills with modern ML stacks (e.g., PyTorch), including debugging and performance/reliability-minded development
Relevant industry experience (typically 5+ years); advanced degrees are valued, but depth of applied experience is important

Nice to Haves

Experience with inference optimization or deploying large models under latency/compute constraints
Experience improving data/training pipelines and working across infrastructure constraints (distributed training, efficiency, reliability)
Proven technical leadership (tech lead ownership, mentoring, setting direction across an area)

Lead and execute Gaia's post-training and closed-loop pipeline, fine-tuning and aligning the world model through post-training experimentation and targeted data curation
Push Gaia's autoregressive generation towards longer, more stable rollouts, and make the model deployment-ready, inference time and reliability included
Contribute to broader model architecture and training-strategy decisions where they intersect with pre- and post-training and the application layer
Partner closely with research, applications, simulation engineering, and cloud/infrastructure teams to translate post-training improvements into measurable downstream impact
Provide technical leadership through mentorship, review, and setting high engineering/research standards

This role is a full-time role based in London, UK (hybrid)
Hybrid working policy that combines time together in offices and workshops and time spent working from home
Core working hours for scheduling flexibility