Senior Computer Vision Engineer — ADAS Perception

Samuel Torres

Contract · Senior · London

This role is no longer accepting applications

Similar open roles you can apply to now:

About the role

This role sits within the Advanced Driver-Assistance Systems (ADAS) perception team of a Tier 1 automotive supplier. We develop perception stacks for production vehicles across three OEM programmes currently in series development. Code written in this team runs in safety-critical embedded environments and is subject to ASPICE Level 2 process requirements. That is not a detail — it shapes how we work. We write requirements before we write code. We conduct peer reviews on every merge. We maintain traceability from safety requirement to test case. If you have worked only in research or startup environments where velocity is the primary metric, this role will feel different. We are looking for a Computer Vision Engineer who can contribute technically to our perception algorithms — primarily camera-based object detection, lane keeping, and sensor fusion — while also operating within a structured development process. The contract is initially 12 months, aligned to the development phase of one of our active programmes. Extension is likely but programme-dependent.

Responsibilities

Develop and optimise camera-based perception algorithms for object detection and lane boundary estimation
Contribute to sensor fusion pipeline development integrating camera and radar inputs
Conduct model evaluation against NCAP and programme-specific performance targets
Produce and review technical documentation in compliance with ASPICE Level 2 requirements
Participate in design and code reviews with the full perception team and contribute to knowledge transfer within the group

Requirements

6+ years of computer vision engineering with at least 2 years in automotive, robotics, or another safety-critical domain
PyTorch for developing and training perception models — experience with model optimisation for embedded deployment is strongly preferred
Python for data pipeline development, model evaluation, and tooling — you should write clean, reviewable code
Computer Vision: object detection, semantic segmentation, depth estimation, and sensor fusion across camera and radar inputs
Docker for reproducible development and validation environments
Kubernetes for distributed training and evaluation pipeline management
Familiarity with ASPICE, ISO 26262, or equivalent safety process frameworks — you do not need to be a functional safety engineer, but you need to understand why the process exists

Benefits

£580 – £720 per day outside IR35 (umbrella or limited company arrangements supported)
12-month initial contract, extension strongly likely for the programme duration
Remote-first with two to four days per quarter at our UK engineering site
Exposure to active series development programmes across multiple OEM platforms
Access to proprietary annotated datasets and GPU compute infrastructure