Founding Research Scientist – Human Attention & Spatial AI
Location: UK‑based, remote‑first (occasional London visits, 1 day a month)
Employment type: Full‑time
Salary: £60,000–£75,000 base
Equity: 0.1%–3.0% options
Visa sponsorship: Not available (UK Graduate Visa accepted)
Role overview
We helping a small startup hire a Research Scientist to work on applied research at the intersection of human attention, perception and spatial AI. Founding style role, join as their 3rd hire!
The role focuses on building models that predict, interpret and simulate how people attend to and interact with complex visual and spatial environments. You will work on real‑world problems that connect machine learning, computer vision and cognitive modelling, with a strong emphasis on hands‑on experimentation and practical impact.
This position is well suited to researchers interested in spatial AI, embodied AI, XR/VR, simulation, automotive systems, robotics or interactive 3D environments. Prior experience across all of these areas is not required.
What you will work on
You will help design and build models that reason about where people look, what they notice and how attention changes with task, context, motion, layout and semantics.
Example work includes:
- Predicting human attention and gaze in 2D and 3D scenes
- Modelling attention during search, navigation, inspection and decision‑making
- Connecting attention with language, goals, objects and scene context using multimodal models
- Inferring human intent or priorities from gaze, movement and interaction data
- Developing evaluation frameworks for human‑like perception
- Adapting foundation models for spatial and human‑centred AI tasks
- Working with synthetic and real‑world data such as images, video, depth, 3D scenes, gaze and interaction logs
- Collaborating with engineers to move research prototypes towards robust systems
Responsibilities
In this role, you will:
- Develop machine learning models for attention, perception, behaviour or intent prediction
- Design experiments, baselines, metrics and validation protocols
- Train and evaluate models using visual, spatial, temporal and multimodal data
- Read, reproduce and extend relevant research
- Work with 2D and 3D data representations including images, video, depth, point clouds and simulated environments
- Communicate research decisions clearly through notes, demos, presentations and code
- Contribute to research direction and technical roadmap
Essential requirements
- PhD in a relevant field (e.g. machine learning, computer vision, robotics, HCI, cognitive science, neuroscience or graphics)
- Strong research track record (top‑tier publications, patents or substantial open‑source work)
- Hands‑on deep learning experience
- Strong Python skills and experience with PyTorch
- Ability to independently design, run and evaluate ML experiments
- Experience working with complex datasets and rigorous evaluation
- Genuine interest in human attention, perception or behaviour
- Clear communication skills and comfort working in a small, fast‑moving team
Useful experience (not required)
- Human attention, gaze modelling, visual saliency or cognitive modelling
- 3D computer vision, spatial AI, simulation, XR/VR or robotics
- Multimodal or temporal modelling
- Egocentric video, eye‑tracking or behavioural data
- Foundation model fine‑tuning or evaluation
- Research engineering and reproducible ML pipelines
You are likely to enjoy this role if you:
- Are curious about how humans perceive and interact with the world
- Are comfortable with ambiguity and open‑ended research problems
- Are technically strong and able to build and debug your own models
- Care about rigour, evaluation and evidence
- Enjoy working in a startup‑style, research‑led environment
- Want to turn research ideas into real systems