Software Engineer, AI Libraries

Overview

In this role, you will help build and evolve Wayve’s AI Platform, enabling ML engineers to train, evaluate and deploy models at scale. You will design modular, scalable Python libraries and improve reliability, performance and developer experience for teams working with large GPU clusters and diverse sensor data. You’ll collaborate with cross-functional teams to translate complex requirements into robust software components. This is a high-impact platform engineering position central to scaling our autonomous-driving technology.

Pay / Benefits

• salary benchmarking
• equity
• relocation support and visa sponsorship
• hybrid working
• learning and development budgets
• health insurance and dental benefits`,`enhanced parental leave and pension where applicable`,`wellbeing partnerships and team socials

Responsibilities

• Design and implement Python-based libraries and tools to support multimodal data loading, distributed training and inference, checkpointing and observability
• Improve reliability, performance and developer experience of systems used across Wayve
• Collaborate with ML teams to understand workflows and translate requirements into scalable software
• Support and enable teams training models on large GPU clusters and working with camera, radar and lidar data sources
• Lead projects from discovery through delivery, and contribute to engineering standards

Key requirements

• extensive Python experience
• designing, delivering and maintaining high-quality software systems at scale
• understanding of software architecture, testing and observability
• experience with concurrent, parallel or distributed computing
• familiarity with ML frameworks such as PyTorch or TensorFlow
• cloud environments, ideally Azure

• collaboration with cross-functional teams
• clear communication and ability to translate requirements into software
• initiative in leading projects from discovery to delivery
• Python
• distributed computing (concurrency/parallelism)
• PyTorch or TensorFlow

…