Waabi, founded by AI visionary Raquel Urtasun, is the leader in Physical AI. With a world-class team, we're unlocking the next era of autonomous transportation with technology that's powering commercial autonomous trucks and robotaxis. Waabi is backed by and partners with world leaders in AI, automotive, logistics, and deep tech.
With offices in Toronto, San Francisco, Dallas, and Pittsburgh, Waabi is growing quickly and looking for diverse, innovative and collaborative candidates who want to impact the world in a positive way. To learn more visit:
www.waabi.ai
As a Research Engineer focused on Generative AI Deployment, you will own the path from research to production for the neural and generative models that power Waabi World, our multi-sensor simulator. You will partner with research teams on models such as feed-forward scene reconstruction models, diffusion models, and vision foundation models. Your job is to make those models production ready: fast, memory-efficient, reproducible, and numerically trustworthy at the scale of closed-loop simulation.
You will…
-
Be part of a multidisciplinary team of Research Scientists and Engineers building a best-in-class multi-sensor simulation stack.
-
Take cutting-edge models from research to production. Turn research code and checkpoints into versioned, reproducible, exportable production models. These models run in closed-loop simulation next to GPU renderers and the autonomy stack, and in the pipelines that reconstruct 4D simulation scenes.
-
Make generative models leaner and faster to run. Profile end to end, from the model graph to GPU kernels and memory movement. Apply techniques like model compilation (e.g., TensorRT), mixed precision and quantization, batching, CUDA graphs, and custom kernels to achieve latency and memory budgets. For productionizing diffusion and flow-matching models, this can also include domain-specific techniques like faster sampling techniques and distillation.
-
Own correctness as part of performance. Build benchmarks and numerical-parity tests focused on simulation realism and downstream autonomy metrics. These checks make sure that each speedup keeps simulation realism intact, and they catch regressions in CI.
-
Build shared deployment tooling. Design the export, versioning, model registry, and build tooling that lets the sensor simulation team ship new models to simulation with confidence.
-
Shape models for deployment. Work with research scientists early, so that new architectures are exportable, efficient, and hardware-friendly from the get-go.
