Here at Humanoid, we believe in a future where robots amplify human potential. That’s why we’ve set out on a mission to build the world’s most capable, commercially-scalable, and safe humanoid robots. We’re bringing that mission to life with HMND‑01 Alpha - our rapidly developed humanoid platform now running in real industrial pilots - and we’re growing the team to take it even further.
We're hiring a strong Manipulation Capabilities Engineer to join our team based in London.
In this role, you will work on teaching our robots to manipulate the world around them.
This is a role at the intersection of applied deep learning and robotics, and to be set up for success you need both experience of working with real robot hardware (e.g. identifying issues in control or teleop), and applied deep learning (you don’t have to be an expert on cutting edge neural network techniques, but you should be perfectly capable of curating data, fine-tuning a policy on that data and hypothesising potential mitigations when something doesn’t work).
Post-train manipulation policies via behaviour cloning and RL; own the full loop from data to deployment.
Come up with data preprocessing strategies to improve the quality of collected data.
Work with the simulation team to set up RL training using digital twin, and then iterate on reward and simulation quality to ensure successful transfer to the real world.
Partner with the data collection organization to drive data collection activities for a specific capability: specify what good data looks like, ensure diversity and coverage, and iterate on instructions.
Expand observation and action spaces with new components required to support novel capabilities, and work with the Teleoperations team to expose these components to robot operators.
Partner with Teleoperations and Controls teams to improve motion smoothness and teleoperation experience.
Interface with hardware design team to ensure that manipulation team findings regarding the current generation of hardware are reflected in future designs.
3+ years working on robots (industry or research) with shipped artifacts to show for it. A good understanding of modern teleoperation and low-level control stack.
Experience with neural network post-training.
Familiarity with deep learning infrastructure: streaming datasets, checkpointing & state management, distributed training, PyTorch or JAX. Ability to profile & debug numerics and write maintainable research code.
Good familiarity with modern software engineering practices.
Ability to document experiments clearly and communicate trade‑offs crisply.
Nice to have:
Experience training VLA models for manipulation (autoregressive, diffusion or flow-matching based). Familiarity with OpenVLA, Physical Intelligence (π) models, or similar open VLA frameworks.
Experience applying RL to robotics problems.
Publications at top-tier robotics or deep learning conferences or equivalent open‑source contributions.
Competitive equity: stock options with meaningful upside as we scale.
30+ days time off, including 23 days annual leave, all UK bank holidays, and additional company closure days (including Christmas–New Year shutdown).
Private healthcare, including virtual and in-person care.
Pension scheme with 8% total contribution (5% employee, 3% employer) on full earnings.
Free daily breakfast, catered lunch, and snacks in-office.
Work at the frontier - collaborate daily with world-class engineers, researchers, and product experts building the next generation of AI and humanoid robotics.
Real ownership - direct access to founding leadership, meaningful input on product direction, and the ability to drive key initiatives from day one.
