Wayve is looking to develop the models, infrastructure, and tooling that power their next-generation synthetic data platform, specifically productizing their foundation model for synthetic multimodal video (GAIA) into scalable tools that generate realistic and controllable sensor data for evaluating their autonomous driving system.
Requirements
- Proven experience developing and deploying machine learning models, ideally involving visual data such as images, video, or 3D scenes
- Strong fundamentals in machine learning, with the ability to reason about model architecture, training dynamics, data requirements, and failure modes
- Practical experience with ML Ops principles, including model versioning, training reproducibility, CI/CD pipelines for ML, monitoring, and observability
- Ability to write clean, efficient, and maintainable code in Python, with a solid understanding of software engineering best practices
- Familiarity with deep learning frameworks such as PyTorch or TensorFlow, and experience building or modifying training and inference pipelines
- Experience training or fine-tuning generative models such as diffusion models, GANs, NeRFs, or other video or view synthesis architectures
- Background in vision, perception, or 3D scene understanding, including experience with temporal or multimodal data (e.g., camera, lidar, radar)
Responsibilities
- Collaborate with researchers to bring cutting-edge architectures into production, adapting experimental models for performance, maintainability, and integration into simulation workflows
- Train, and improve generative models that produce realistic and controllable multimodal sensor data, contributing directly to how we evaluate and validate our autonomous driving system
- Build scalable and efficient pipelines for inference and evaluation of large generative models on real and synthetic visual data
- Apply ML Ops best practices: reproducibility, model versioning, evaluation pipelines, and deployment hygiene, to ensure our models can operate reliably in production environments
- Develop tools to monitor, measure, and improve model quality, generation throughput, modality consistency, and domain coverage
- Write clean, modular, and testable code that interfaces well with internal simulation platforms, sensor emulators, and evaluation systems
- Engage in technical design discussions, participate in design reviews, and help shape architecture choices across model and infrastructure layers
Other
- Comfortable working in collaborative, cross-functional teams and contributing to shared goals across engineering, research, and product groups
- Curious, pragmatic, and capable of diving into unfamiliar code, tools, or domains to solve high-impact problems
- Uphold a culture of engineering rigor and continuous learning—through mentoring, code reviews, shared experiments, and thoughtful documentation
- Work closely with teammates across ML, simulation, cloud, and autonomy to ensure our outputs are aligned with real-world system needs and contribute to tangible performance improvements
- Hybrid working policy that combines time together in our offices and workshops to fuel innovation, culture, relationships and learning, and time spent working from home.
