Чтобы адаптировать резюме под вакансию или составить сопроводительное письмо, загрузи резюме
описание
Complete the challenge above and submit your solution as a public GitHub repository by Friday, 9 October 2026, 23:59 BST. Include a README/Presentation with instructions to run your system, example outputs, and a note on your design choices, what worked and what didn’t. The data you personally collected must play a role in the approach.
Humanoid is developing commercially scalable, safe humanoid robots, including the HMND-01 platform for deployment in real industrial environments. It builds software systems that enable robots to operate effectively in the real world and expand human capabilities.
задачи
Train language-vision conditioned manipulation policies using reinforcement learning in the real world
Construct challenging and diverse manipulation task suites and reinforcement learning models in simulation using Isaac Sim and MuJoCo
Experiment with transferring policies trained in simulation to the real world
Develop action-conditioned video prediction and physically consistent dynamics models over long horizons
Use world models as learned simulators to score candidate policies offline and generate synthetic rollouts for training
Build fidelity metrics that quantify where world models can be trusted
Explore in-context learning, short- and long-term memory, and post-train VLA models for production-grade use cases
Work with different data modalities, close the embodiment gap between human and robot data, and improve data diversity and attribution
Optimise models for real-time edge inference on robot hardware, including profiling, quantisation, and latency/throughput trade-offs
Improve training and data-loading performance across distributed GPU infrastructure
требования
Pursuing or holding a master’s degree or PhD in computer science, machine learning, robotics, or a related field
Strong foundations in machine learning
Strong Python skills and hands-on experience with PyTorch or JAX
Interest in one or more of reinforcement learning, world models and generative video, VLA/multimodal models, or ML systems and inference optimisation
Experience running experiments and interpreting results rigorously
Ability to take ownership and iterate with guidance
Strong problem-solving skills and attention to detail
Ability to learn quickly and work in a research-driven, fast-moving environment
условия
Internship for 12 to 24 weeks, 5 days per week
Flexible start date
Competitive pay and perks
Free daily breakfast, catered lunch, and snacks in the office
Daily collaboration with engineers, researchers, and product experts building AI and humanoid robotics
Direct access to founding leadership, input on product direction, and the ability to drive initiatives from day one
.svg)
