10 June 2024  |  Leadership

Science at Wayve: Embark on the Journey of a Lifetime

Discover what sets Science at Wayve apart as we explore our innovative approach to driving rapid technological progress and personal growth. We hope this inspires you to consider a rewarding career with us at Wayve.

Jamie Shotton Chief Scientist

Wayve started in a Cambridge garage in 2017 with a handful of scientists exploring how end-to-end deep learning could be used to drive a car. Fast forward to 2024: Wayve has raised over $1B, grown to 300 people, and established offices in 3 major AI hubs. We’re on a bold, exciting mission to reimagine autonomous mobility through Embodied AI. Our world-class Science team, a key part of our past and future success, is expanding across London, Silicon Valley, and Vancouver to accelerate our mission.

In this blog, I offer a glimpse into what sets Wayve apart: why we’re tackling the right problem, our fantastic team, and how we work in Science to drive rapid progress, learning, and opportunities for personal growth. We hope this will inspire you to consider a career at Wayve.

My journey to Wayve

My passion and fascination for technology began at a young age. My parents tell the story of a hurried hotel departure after they found three-year-old me halfway through disassembling the room’s television! Today, my fascination extends beyond understanding how things work to exploring how new technologies can improve people’s lives.

I joined Wayve three years ago, knowing that the AI revolution was beginning and that enormous opportunities for positive impact were emerging in Embodied AI. Joining Wayve was a leap into a future I wanted to help create—a future where autonomous vehicles and robots enhance our everyday lives and improve safety for all road users. Bringing Embodied AI out of the lab to safely operate around people in the unconstrained complexity of the real world is one of the hardest and most interesting challenges of our time. At Wayve, I’ve found a place where our multidisciplinary team collaborates deeply to make this vision a reality day by day.

It’s an incredible time to be at Wayve.

Wayve stands at a pivotal moment: we have everything we need to focus on the interesting, hard problems while still allowing individuals and small teams to have a mammoth impact across the company and industry.

We’ve moved beyond the early-stage startup grind of building basic infrastructure, securing enough capital to stay afloat, and finding enough compute and data resources to make a meaningful dent in the problem. Through grit, determination, and a fair sprinkling of luck, we’ve beaten the 100-1 odds against early-stage deep tech startups. We’ve developed an industry-redefining technology foundation, secured a huge Series C fundraise, built an incredibly cohesive team, established deep industry partnerships on world-class compute and data resources, and have the passion and determination to go the distance to ship transformative Embodied AI products that solve this problem at scale.

Despite our growth, we remain fast-moving and nimble. Unlike big tech, our culture allows individuals and small teams to undertake significant projects that have company- and industry-shaping impact, as demonstrated by our GAIA and LINGO initiatives last year. We are a highly mission-driven group united by a singular vision, free from the red tape and politics that slow larger companies.

Science: Unlock, Enable, Disrupt

Research has been in Wayve’s DNA from its inception in Cambridge, with early work such as Learning to Drive in a Day and Dreaming about Driving and recent work such as GAIA and LINGO. We have always acknowledged the immense challenge—and opportunity—ahead and the critical importance of investing in groundbreaking research and innovation. To maintain this focus as the company scales, we have spent the last few years building out our Science division.

“Our current technology is market-leading today, but we aim to stay ahead and enable autonomy everywhere. For this reason, perhaps unusually for a startup, we have invested in a Science team from the beginning, balancing our ability to ship week-to-week improvements by innovating iteratively, with long-term innovation that nurtures 10x ideas that may take years to mature.”

Alex Kendall, Co-founder and CEO

Our Science team, a global group of researchers and engineers, unlocks disruptive innovation to solve self-driving and enable Embodied AI. With a distinct culture, rhythm, and goals, we aim to:

  • Unlock breakthrough capabilities through cutting-edge research that redefines the state of the art in Embodied AI.
  • Enable and mature new technologies into our production stack in partnership with our engineering teams.
  • Disrupt ourselves and the industry before someone else does.

We built the Science division to empower the world’s leading Embodied AI researchers and engineers to do their greatest work and deliver step-change advances for the field and our company.

We’ve instilled a strong culture of scientific exploration and rigour. Our team thrives on reaching for the stars with our ambitions, constantly challenging assumptions, and fostering both individual and collective growth through continuous learning and knowledge sharing. We allow ideas the time they need to mature while keeping a strong connection with the rest of the company to ensure a smooth and rapid integration of new ideas into production.

Our scientists and engineers have access to world-leading resources, allowing us to expand the frontiers of Embodied AI. Data and compute are the lifeblood of our company. Our internal fleet operations, partner relationships, and licensing agreements provide us with an exponentially growing corpus of diverse training data for driving and beyond. Our deep partnership with Microsoft has enabled a world-class GPU training platform, supplying the compute power necessary for transformative research.

We value self-driven individuals who take the initiative, iterate quickly, and accelerate their teams. Experimentation is fast and easy at Wayve, allowing you to see your work’s impact almost immediately. Perhaps uniquely in the world of Embodied AI, new joiners can test their ideas on the road within days or even hours of starting.

“At Wayve, we have access to ever-growing datasets and partnerships that provide industry-leading compute power. Our researchers are advancing state-of-the-art research in generative simulators, integrating language with robotics, developing foundation models for generalized Embodied AI, and more, redefining what is possible in autonomous driving.”

Cameron Tuckerman-Lee, Machine Learning Engineer

“Conducting real-world experiments on vehicles to safely test models on public roads in London and witnessing the performance improvements with new techniques has been fantastic. What I enjoy most about working at Wayve is how the scientific method is central to our process, from the way we measure performance improvements to how we pursue increased transparency in decision-making.”

Ana-Maria Marcu, Machine Learning Engineer

Our Team

It may be a cliché, but our people truly are our greatest asset. Our Science team is filled with inspiring individuals who open your eyes to new possibilities and challenge you to achieve your best. Our researchers and engineers come from diverse backgrounds in industry and academia, which fuels our creativity and drives success. Our team includes veterans from big tech companies like Microsoft, Google, Amazon, and Meta, as well as startups focused on autonomous driving, generative AI, robotics, mixed reality, sports analytics, and more. Many have Master’s and PhDs in machine learning, computer vision, robotics, neuroscience, and related fields from top institutions such as Cambridge, Berkeley, UCL, TUM, Stanford, EPFL, Oxford, and CMU.

We also value those who have taken different paths and appreciate their unique perspectives and contributions. We are committed to an inclusive environment and have excellent managers dedicated to helping you achieve your full potential.

Our team thrives in offices in three exciting locations: London, Silicon Valley, and Vancouver. Each place has its own character and focus, but we collaborate across geographies to advance our mission. We also welcome several interns each year, giving them a chance to deeply embed themselves in our teams, explore new and exciting directions, and experience what a full-time position at Wayve would be like.

“After years of chasing conference deadlines in a seemingly never-ending cycle of incremental papers, Wayve has given me the opportunity to work on a world-changing problem in a vibrant environment. Training a driving model and seeing it on the road, in the heart of London the next day is a truly unique experience.”

Remi Tachet des Combes, Applied Scientist

“At Wayve, you can train and deploy driving models, experience them firsthand in a car, and explore technologies like World Models, LLMs, NeRF, and RL. Science at Wayve satisfies all your curiosities about Embodied AI and pushes the frontiers further every day.”

Long Chen, Applied Scientist

What we focus on

We tackle a wide range of the hardest technical problems in Embodied AI. We don’t organise our work by the traditional AV1.0 categories of perception, prediction, and planning—distinctions that are largely meaningless in AV2.0. Instead, we build centres of excellence around broad themes such as behaviour, representation learning, language, graphics, and vision.

Our technical leaders initiate and drive cross-functional projects, which we resource flexibly as we progress and learn. These projects aim to improve our on-road driving performance, accelerate in-silico measurement accuracy, develop core foundation models that underpin our AI, inspire unique product offerings, increase the safety and explainability of our models, and achieve fundamental technical breakthroughs. Today, we have projects focused on Embodied AI foundation models, safety architectures, vision-language-action grounding, neural dynamic scene reconstruction, offline RL, imitation learning, training and inference efficiency, explainability, and more.

Wayve is focused on bringing our first driving products to market. However, we also see significant future opportunities in the broader Embodied AI space. Trends like cross-embodiment foundation models (e.g. RT-X) show promise for generalisation across multiple robotic platforms. We are positioning ourselves to benefit from these trends through both data and research investments.

This is real-world driving footage of a LINGO-2 test run. The text is output from the LINGO-2 model and overlaid on the video.
GAIA-1, our generative world model, can create multiple weather conditions from text prompts.
Our closed-loop data-driven neural simulator, Ghost Gym, uses neural rendering to create photorealistic 4D worlds for training.

How we work

Each project tackles a bold, specific hypothesis with a clear “what, why, and how”, reduces this to practice, and validates it. We prioritise simplicity in solutions and celebrate effective results over novelty. We sweat the details and visualise everything, which is crucial for progress in machine learning. We love our data – its quality, diversity, and quantity – just as much as our model architectures. Staying connected to the problem, we regularly take on-road demo rides to experience firsthand what we’re building.

Major research breakthroughs often come when you join the dots between seemingly unrelated topics. Our weekly internal Science Forum combines deep technical presentations of the latest work from across the team and a powerful interactive Q&A session to rapidly share ideas and ask questions. We stay up to date with the wider advances in the community through guest speakers, reading groups, and conference attendance and move fast to integrate new ideas within days of them being released, iterating rapidly without hesitation to abandon dead-ends.

Case Study: World Modelling

From dreaming about driving (2018) to future prediction in camera perspective (2020) and birds-eye view (2021) and combining future prediction and behavioural cloning (2022), world models have consistently played a role in Wayve’s research. Each project began with the insight that understanding the environment would benefit our driving intelligence, reinforcing our belief that world models are fundamental to our technology. In late 2022, we launched GAIA, our most ambitious world modelling project to date, leveraging the power of self-supervision to build a multi-modal world model capable of impressive video generative capabilities.

“Embarking on this ten-month journey, we were optimistic and ready to execute our ambitious plan but aware of the magnitude of the challenges ahead. As a team, we faced many hurdles, brainstorming ideas and enhancing efficiency to scale GAIA from 100 million parameters to 9 billion parameters. We celebrated many milestones, including GAIA’s first generated video and its ability to create out-of-distribution states, such as zig-zagging through traffic, driving in the snow in London, and generating realistic birds in the sky. These achievements highlighted GAIA’s enhanced realism and captivating capabilities as it scaled.”

Gianluca Corrado, Applied Scientist

Our work often has broader fundamental relevance to the AI community. We don’t chase the next conference deadline but publish when we have significant, impactful advances to share. We further contribute to the academic community through workshops, tutorials, summer schools, and dataset releases.

To accelerate our work and maximise its downstream impact, we constantly collaborate closely with other teams across Wayve, including compute platform, AI production, data partnerships, and public policy. We stay connected to leadership through regular listening tours, all-hands meetings, and Q&A sessions.

Achieving amazing results and downstream impact requires hard work and a passion for problem-solving. We also value balance and clear boundaries for personal well-being and to foster creativity. Working smart is as important as working hard: we focus on solving the right problems, aiming for meaningful results, and constantly adjusting course to find the shortest path to solutions.

Ultimately, the work should be intrinsically fun and fulfilling, helping you build technical mastery and rewarding relationships. Our goal is to make daily progress, knowing that these small steps compound and lead to exponential advances over time.

A Culture That Empowers You To Be Your Best

I’m proud of our culture. It’s not just about socials, after-work sports, or office design – though those matter. Culture is how we work together and what we value:

  • Growth mindset. We embrace continuous learning and self-improvement, recognising that we may not know everything, yet, but we know how to learn and grow.
  • Ambition without fear of failure. We pursue work with huge potential and understand that when you aim high, not everything will succeed as planned – that’s the nature of research. We strive to execute strongly and learn from every experience.
  • Passion for solving the unsolvable. We thrive on tackling hard, ambiguous, and complex problems with optimism.
  • Commitment to excellence. We hold ourselves and each other accountable to being the best at what we do. We set high standards, are rigorous in our work, and always push to raise the bar. We recognise and reward high-performing individuals and teams. Wayve is where you come to do the best work of your career.
  • Humble collaboration. Building large-scale learning systems that power modern AI is all about teamwork. Leave your ego at the door.
  • Diverse and inclusive. We listen to and value the diverse perspectives of everyone on the team. Our people join Wayve from all over the world, including Sweden, Canada, Germany, Japan, Romania, the US, Italy, India, and the UK. We welcome new joiners warmly and foster lifelong friendships.
  • Curiosity. As true scientists, we are always asking why and striving to uncover what makes things tick.
  • Fast iteration. We move quickly to get things done, go from working to working, and don’t let perfectionism and bureaucracy get in our way.
  • Constructive candour. Even when it feels hard to do, we give and receive feedback that helps us grow and solve problems faster.
  • Open communication. Transparency improves feedback cycles and provides visibility at all levels of the organisation.
  • Persistence and grit. The journey is long and littered with obstacles. We persist, knowing we can’t lose if we don’t quit.
  • Data-driven. We use data not just to train our models but also to make informed decisions.
  • Autonomy and empowerment. Our technical leaders are empowered to choose their battles and drive outcomes for the company with the resources they need.

“Science at Wayve fosters a culture where good ideas can come from anywhere, at any time. We share failures and learnings frequently and publicly, celebrating them alongside breakthroughs. We stay nimble, constantly trying new things, and move on quickly. The only constant is change, which keeps things interesting! My career at Wayve has been a period of immense learning—filled with new domains, growth opportunities, and a research culture that fuels my curiosity.”

Sofia Josefina Lago Dudas, Applied Scientist

Join Us on This Remarkable Journey

Wayve is the place where we dream big and turn those dreams into reality. We’re not just building self-driving cars but creating a safer, more sustainable, and connected future. We’re establishing the foundations of Embodied AI to improve billions of lives and drive trillions in economic growth across the world.

We seek the best and brightest people passionate about making a lasting positive impact on the world. Wayve will challenge you to excel in a team that values innovation and collaboration, helping you learn and grow faster than you thought possible. 

If you’re ready for the journey of a lifetime and want to be part of a team redefining an industry, we’d love to hear from you. Check out our open roles or contact our Science recruitment team.

Back to top