Transform lives by leading backend development for Wispy's AI-power augmented reality system at CrossSense.
Overview
£49920 - £62400
London/ Remote
Expires at anytime
The mission
CrossSense is a breakthrough AI-powered application for augmented reality smartglasses designed to help people with dementia maintain their autonomy. Because the technology sees what the user sees and hears what they hear, our AI companion, Wispy, provides real-time guidance, a "cognitive scaffold", to assist with memory gaps, task execution, and social interaction.
We agree with one of the experts by experience who told us that life without joy is not worth living. CrossSense isn't about managing decline; it's about enabling people to live fulfilling lives on their own terms. The system acts as a "missing piece" that bridges the gap between what a person intends to do and what the condition disrupts. Rather than performing tasks for the user, we support their existing cognitive resources to maintain functional independence and dignity.
We have been developing extended reality software since 2016. We are finalists for the Longitude Prize on Dementia, with R&D funding from NIHR and Innovate UK. This summer we begin our third research study, with commercial release planned for late 2026.
The role
You'll lead backend development for Wispy's cloud infrastructure, the "brain" that processes real-time audio, video, and sensor data from thin clients (AR glasses and smartphones) hosted on European servers.
This is a hands-on technical leadership role. You'll architect systems, write code, and shape how we build software. We use AI-assisted development and are exploring how human-AI collaboration can amplify developer capability without sacrificing learning, fun and code quality.
What you'll ship
- Real-time streaming pipelines for audio and camera feeds with very low latency
- A dual-layer knowledge graph (Graphiti + Neo4j) storing evidence-based guidance alongside personalised user routines
- Multimodal AI agent orchestration integrating speech-to-speech models (Gemini Live), computer vision (YOLOE, Depth Anything), and voice analysis (SenseVoice)
- Privacy-by-design systems that store semantic facts rather than raw data
Key responsibilities
- Design and scale backend infrastructure on Kubernetes with horizontal scaling
- Evolve the knowledge graph architecture for personalised, temporal user data
- Orchestrate multimodal AI agents using frameworks like Google ADK
- Implement privacy-preserving data pipelines for regulated healthcare contexts
- Collaborate with the AR client team to ensure interoperability across OpenXR devices
Required experience
- Strong backend development in Python with async patterns
- Cloud-native architecture with Kubernetes and container orchestration
- Graph databases (Neo4j) and complex data modelling
- Real-time protocols (WebSocket, MessagePack)
- Experience with agentic AI frameworks (Google ADK, LangGraph, or similar)
- Deploying and orchestrating ML pipelines spanning computer vision and speech models
- Commitment to data ethics and secure handling in regulated environments
Nice to have
- Unity/C# for occasional AR client support
- Healthcare, clinical research, or assistive technology experience
- Accessibility standards (WCAG) or designing for cognitive accessibility
- Temporal data modelling and event sourcing patterns
- Observability tooling (OpenTelemetry, distributed tracing)
- Hexagonal architecture and dependency injection patterns
What we offer
- Salary: £49,920 (4-day week) to £62,400 (5-day week) - will increase as we raise finance
- Equity: 1% stake plus future EMI share options
- Flexibility: Open to a 4-day working week
- Location: Hybrid, London office ~2 days/week
- Benefits: Workplace pension, flexible hours, learning budget, home office support.
- Impact: Directly shape a product that supports people living with dementia, their loved ones and communities.
- Contract type: Permanent with 6 month probation period.
How we work
We believe it's never too late to pursue meaningful goals. We think long-term while acting in the moment. We work as a joint effort, supporting each other's potential, just as CrossSense empowers user agency to be independent.
We make room for questioning, uncertainty, and diverse perspectives. We prioritise kindness and shared benefits over competition. We honour the trust placed in us by the 250+ people with lived experience of dementia who have shaped CrossSense through co-design and consultation: their unmet needs, privacy and dignity guide our trajectory.
Start date
Late February
Application process
We'll interview selected candidates on a rolling basis.
How to apply
Send your CV and a short note on what draws you to this work. We'd love to hear about backend systems you've built and any experience with real-time or AI-adjacent projects.