AI Research Scientist

Overview

Overview:

At Zebra, we are a community of innovators who come together to create new ways of working. United by curiosity and a culture of caring, we develop smart solutions that anticipate our customer’s and partner’s needs and solve their challenges.

Being part of Zebra Nation means you are seen, heard, valued, and respected. Drawing from our unique perspectives, we collaborate to deliver on our purpose. Here you are part of a team pushing boundaries today to redefine the work of tomorrow for organizations, their employees, and those they serve.

You’ll have opportunities to learn and lead in a forward-thinking environment, defining your path to a fulfilling career while channeling your skills toward causes you care about—locally and globally.

Come make an impact every day at Zebra.

What We're Looking For:

We are looking for an AI Research Scientist specialising in SLAM, Visual-Inertial Odometry (VIO), and 3D Spatial Perception to join our central CTO Group. In this horizontal, high-impact organisation, you will collaborate with cross-functional business units across the globe and work closely with our London-based research hub of ~15 AI specialists.

In this role, you will be bridging the gap between innovation and on-device execution -prototyping new algorithms, training spatial models, and optimising them to run directly on resource-constrained mobile edge devices to power next-generation Augmented Reality (AR), enterprise vision, and edge intelligence.

You will be based at our central London office (Endeavour House, Shaftesbury Avenue, WC2H 8JR) on a flexible hybrid model - collaborating in person 3 days a week, with 2 days working from home.

Responsibilities

• Research, design, and implement robust Visual-Inertial Odometry (VIO), 6-DoF camera pose estimation, and monocular SLAM pipelines tailored for resource-constrained edge hardware.
• Develop advanced algorithms for spatial localisation, multi-view ray triangulation, sensor fusion (IMU + Camera), and plane-fitting (e.g., RANSAC, PnP) to achieve sub-centimetre overlay accuracy and eliminate tracking drift.
• Translate research papers into real-time algorithms, optimising neural networks and spatial pipelines for low-latency execution on mobile NPUs, DSPs, and GPUs.
• Build spatial foundations that deliver seamless on-device AR experiences, barcode/shelf product recognition, OCR, and multi-modal edge AI capabilities without relying on cloud connectivity.
• Partner with global hardware, software, and product engineering teams to integrate your models into commercial product roadmaps.