Job summary Autonomous Labs is building the next generation of AI-driven scientific laboratories, with the mission of automating the entire scientific discovery workflow — from experimental design and execution to analysis and iterative learning. Foundation models are a core component of this vision, enabling intelligent, adaptive, and autonomous scientific experimentation. As a member of the technical team focused on foundation model performance, you will play a key role in understanding, evaluating, and improving the capabilities of frontier foundation models for embodied AI and autonomous scientific experimentation. This spans both autoregressive foundation models — the sequence-modelling backbone behind language, vision-language and action prediction — and world models and world action models that learn the dynamics of the laboratory well enough to simulate, plan and act on experiments before they are run. Your research will systematically investigate how training data, model development stages, and training strategies interact to determine model Internal Updated: August 2026 Job description — Internal capability. By uncovering these interactions, you will identify performance bottlenecks and develop novel approaches, such as new reward models, new learning curricula or data mixing strategies, that continuously improve foundation model performance. This role offers a unique opportunity to conduct frontier research at the intersection of large-scale foundation models, data-centric AI, and embodied intelligence while solving real-world scientific problems. You will work closely with other AI researchers, software engineers, robotics engineers, and domain scientists to translate advances in autoregressive foundation models and world models into measurable improvements in autonomous laboratory performance. The successful candidate will design and execute systematic experimental studies to understand how different data mixtures, data quality, model architectures, and training stages influence downstream capabilities and final performance in an end-to-end scientific workflow. You will identify performance bottlenecks and conduct research on improving foundation model performance through a deeper understanding of the interactions between data and models. Requirements Key responsibilities Drive technical work on improving the performance of frontier foundation models — both autoregressive foundation models and world models — for autonomous scientific experimentation. Conduct research into the interactions between training data and foundation models to identify the key factors limiting model performance, and develop novel data-centric approaches for continuous improvement. Define and source the data needed to move model performance, working closely with domain experts across the laboratory to specify, collect, and curate high-value experimental datasets. Design and maintain scalable experimentation pipelines for model training, evaluation, benchmarking, and reproducible research. Analyse experimental results using rigorous scientific methodologies and translate insights into actionable improvements for model performance in a data-centric way. Collaborate closely with AI researchers, software engineers, robotics engineers, and scientific domain experts to ensure research findings translate into impactful real-world scientific capabilities. Contribute novel research ideas and publish high-quality research where appropriate, while maintaining a strong focus on practical deployment. Communicate experimental findings and technical insights clearly across interdisciplinary teams to help shape the future direction of the AutoLab AI platform. Essential knowledge, skills, and experience MSc, PhD, or equivalent industry experience in Computer Science, Artificial Intelligence, Machine Learning, Robotics, or a related discipline. Hands-on experience with frontier foundation models, including autoregressive foundation models such as Large Language Models (LLMs), Vision-Language Models (VLMs) and Vision-Language-Action (VLA) models, as well as embodied AI models and world models. Experience with continuous pre-training, post-training, supervised fine-tuning, reinforcement learning, and other foundation model optimisation techniques. Hands-on experience building end-to-end machine learning pipelines, including data curation, data mixing, model training, evaluation, performance analysis, and iterative model improvement. Experience working within interdisciplinary teams. Excellent written and verbal communication skills, with the ability to communicate complex experimental findings clearly across disciplines. An impact-driven mindset with a passion for continuously improving model performance for real world deployment. Desirable knowledge, skills, and experience In rough order of desirability. Research experience in embodied AI and robotics. Research or engineering experience with world models, model-based reinforcement learning, or learned simulators. Experience in working with scientific domain experts. Experience with large-scale distributed training infrastructure. Publications at leading AI conferences (e.g. NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, CoRL, RSS, or equivalent). Experience working with large multimodal datasets, robotics datasets, or scientific datasets. Experience building scalable and reproducible experimentation and evaluation frameworks. Personal attributes Curious and analytical, with a passion for understanding why foundation models succeed or fail. Strong scientific thinking combined with practical engineering skills. Comfortable working in an iterative, fast-paced research environment. Collaborative and proactive, with the ability to work effectively across AI, robotics, software engineering, and scientific disciplines. Organised, resourceful, and capable of managing multiple experimental projects simultaneously. Mission-driven, with a desire to accelerate scientific discovery through